3 Commits

Author SHA1 Message Date
Rutwik Choughule
a78fba92df update sample cudaNvSci 2022-01-25 17:06:02 +05:30
Rutwik Choughule
876266c941 update Makefiles 2022-01-25 14:06:23 +05:30
Rutwik Choughule
97bb3938a6 update Makefile for matrixMulDrv 2022-01-18 18:55:51 +05:30
4643 changed files with 450755 additions and 214074 deletions

View File

@@ -1,49 +0,0 @@
---
AccessModifierOffset: -4
AlignAfterOpenBracket: Align
AlignConsecutiveAssignments: Consecutive
AlignConsecutiveDeclarations: Consecutive
AlignConsecutiveMacros: Consecutive
AlignEscapedNewlines: Left
AlignOperands: AlignAfterOperator
AlignTrailingComments: true
AllowAllParametersOfDeclarationOnNextLine: false
BinPackArguments: false
BinPackParameters: false
BraceWrapping:
AfterClass: true
AfterControlStatement: false
AfterExternBlock: true
AfterFunction: true
AfterStruct: true
AfterUnion: true
BeforeCatch: true
BeforeElse: true
IndentBraces: false
BreakBeforeBraces: Custom
BreakBeforeConceptDeclarations: true
BreakBeforeBinaryOperators: NonAssignment
BreakBeforeTernaryOperators: true
BreakConstructorInitializers: BeforeComma
BreakInheritanceList: BeforeComma
ColumnLimit: 120
DerivePointerAlignment: false
FixNamespaceComments: true
IncludeCategories:
- Regex: '^<.*>'
Priority: 1
- Regex: '^".*"'
Priority: 2
SortIncludes: true
IncludeBlocks: Regroup
IndentWidth: 4
MaxEmptyLinesToKeep: 2
PointerAlignment: Right
SortUsingDeclarations: true
SpaceAfterCStyleCast: false
SpaceBeforeAssignmentOperators: true
SpaceBeforeParens: ControlStatements
Standard: c++17
TabWidth: 4
UseTab: Never
...

6
.gitignore vendored
View File

@@ -1,6 +0,0 @@
build
.vs
.clangd
test
settings.json
launch.json

View File

@@ -1,116 +0,0 @@
# Copyright (c) 2024, NVIDIA CORPORATION.
ci:
autofix_commit_msg: |
[pre-commit.ci] auto code formatting
autofix_prs: false
autoupdate_branch: ''
autoupdate_commit_msg: '[pre-commit.ci] pre-commit autoupdate'
autoupdate_schedule: quarterly
skip: []
submodules: false
repos:
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v5.0.0
hooks:
- id: end-of-file-fixer
exclude: |
(?x)^(
.*\.raw$|
.*\.bin$|
.*\.dat$|
.*\.nv12$|
data/.*|
Common/.*
)
files: |
(?x)^(
.*\.txt$|
.*\.md$|
.*\.cpp$|
.*\.cxx$|
.*\.hpp$|
.*\.h$|
.*\.cu$|
.*\.cuh$|
.*\.py$|
.*\.json$
)
- id: mixed-line-ending
exclude: |
(?x)^(
.*\.raw$|
.*\.bin$|
.*\.dat$|
.*\.nv12$|
data/.*|
Common/.*
)
files: |
(?x)^(
.*\.txt$|
.*\.md$|
.*\.cpp$|
.*\.cxx$|
.*\.hpp$|
.*\.h$|
.*\.cu$|
.*\.cuh$|
.*\.py$|
.*\.json$
)
- id: trailing-whitespace
exclude: |
(?x)^(
.*\.raw$|
.*\.bin$|
.*\.dat$|
.*\.nv12$|
data/.*|
Common/.*
)
files: |
(?x)^(
.*\.txt$|
.*\.md$|
.*\.cpp$|
.*\.cxx$|
.*\.hpp$|
.*\.h$|
.*\.cu$|
.*\.cuh$|
.*\.py$|
.*\.json$
)
- repo: https://github.com/pre-commit/mirrors-clang-format
rev: v19.1.6
hooks:
- id: clang-format
types_or: [file]
files: |
(?x)^(
^.*\.c$|
^.*\.cpp$|
^.*\.cu$|
^.*\.cuh$|
^.*\.cxx$|
^.*\.h$|
^.*\.hpp$|
^.*\.inl$|
^.*\.mm$
)
exclude: |
(?x)^(
Common/.*
)
args: ["-fallback-style=none", "-style=file", "-i"]
- repo: https://github.com/astral-sh/ruff-pre-commit
rev: v0.8.4
hooks:
- id: ruff
name: ruff lint
args: [--fix]
files: ^python/.*\.py$
- id: ruff-format
name: ruff format
files: ^python/.*\.py$

View File

@@ -1,287 +0,0 @@
## Changelog
### CUDA 13.2 (update)
* Added **CUDA Python samples** under `python/`. These scripts use [CUDA Python](https://nvidia.github.io/cuda-python/) (including `cuda.core`) and are organized like the C++ tree: `1_GettingStarted`, `2_CoreConcepts`, `3_FrameworkInterop`, and `4_DistributedComputing`, plus shared helpers in `python/Utilities`. Each sample includes a `README.md` and `requirements.txt`. They are **not** built by the root CMake project; install dependencies with `pip install -r requirements.txt` in the sample directory, then run the corresponding `.py` file as documented in that sample’s README.
* Renamed top-level `Samples` directory to `cpp` to accommodate Python samples alongside existing C++ samples; updated path references in `CMakeLists.txt`, `README.md`, and `Common` headers accordingly.
### CUDA 13.2
* Added the MSVC compile flag `-Xcompiler=/Zc:preprocessor` in CMakeLists.txt to comply with CUDA13.2 CCCL. Previously, using the traditional preprocessor triggered the warning “MSVC/cl.exe with traditional preprocessor is used…”, which now leads to a build error.
### CUDA 13.1
* Minor bug fixes and enhancements, no structural or functional changes
### CUDA 13.0
* Updated the samples using the cudaDeviceProp fields which are deprecated and removed in CUDA 13.0, replacing the fields with the equivalents in "cudaDeviceGetAttribute":
* Deprecated "cudaDeviceProp" fields
`int clockRate; // - Replaced with "cudaDevAttrClockRate"`
`int deviceOverlap; // - Replaced with "cudaDevAttrGpuOverlap */`
`int kernelExecTimeoutEnabled; // - Replaced with "cudaDevAttrKernelExecTimeout`
`int computeMode; // - Replaced with "cudaDevAttrComputeMode" */`
`int memoryClockRate; // - Replaced with "cudaDevAttrMemoryClockRate"`
`int cooperativeMultiDeviceLaunch; // - Deprecated, cudaLaunchCooperativeKernelMultiDevice is deprecated.`
* `0_Introduction`
* `UnifiedMemoryStreams`
* `simpleHyperQ`
* `simpleIPC`
* `simpleMultiCopy`
* `systemWideAtomics`
* `1_Utilitie`
* `deviceQuery`
* `2_Concepts_and_Techniques`
* `streamOrderedAllocationIPC`
* `4_CUDA_Libraries`
* `simpleCUBLASXT`
* `5_Domain_Specific`
* `simpleVulkan`
* `vulkanImageCUDA`
* Updated the samples using the CUDA driver API "cuCtxCreate" with adding the parameter "CUctxCreateParams" as "cuCtxCreate" is updated to "cuCtxCreate_v4" by default in CUDA 13.0:
* `Common`
* `nvrtc_helper.h`
* `0_Introduction`
* `UnifiedMemoryStreams`
* `matrixMulDrv`
* `simpleTextureDrv`
* `vectorAddDrv`
* `vectorAddMMAP`
* `2_Concepts_and_Techniques`
* `EGLStream_CUDA_CrossGPU`
* `EGLStream_CUDA_Interop`
* `threadMigration`
* `3_CUDA_Features`
* `graphMemoryFootprint`
* `memMapIPCDrv`
* `4_CUDA_Libraries`
* `jitLto`
* `7_libNVVM`
* `cuda-c-linking`
* `device-side-launch`
* `simple`
* `uvmlite`
* `8_Platform_Specific/Tegra`
* `EGLSync_CUDAEvent_Interop`
* Updated the sample using CUDA API "cudaGraphAddNode"/"cudaStreamGetCaptureInfo" with adding "cudaGraphEdgeData" pointer parameter as they are updated to "cudaGraphAddNode_v2"/"cudaStreamGetCaptureInfo_v3" by default in CUDA 13.0:
* `3_CUDA_Features`
* `graphConditionalNodes`
* Updated the samples using CUDA API "cudaMemAdvise"/"cudaMemPrefetchAsync" with changing the parameter "int device" to "cudaMemLocation location" as they are updated to "cudaMemAdvise_v2"/"cudaMemPrefetchAsyn_v2" by default in CUDA 13.0.
* `4_CUDA_Libraries`
* `conjugateGradientMultiDeviceCG`
* `6_Performance`
* `UnifiedMemoryPerf`
* Replaced "thrust::identity<uint>()" with "cuda::std::identity()" as it is deprecated in CUDA 13.0.
* `2_Concepts_and_Techniques`
* `segmentationTreeThrust`
* Updated the the headers file and samples for CUFFT error codes update.
* Deprecated CUFFT errors:
* `CUFFT_INCOMPLETE_PARAMETER_LIST`
* `CUFFT_PARSE_ERROR`
* `CUFFT_LICENSE_ERROR`
* New added CUFFT errors:
* `CUFFT_MISSING_DEPENDENCY`
* `CUFFT_NVRTC_FAILURE`
* `CUFFT_NVJITLINK_FAILURE`
* `CUFFT_NVSHMEM_FAILURE`
* Header files and samples that are related with this change:
* `Common/helper_cuda.h`
* `4_CUDA_Libraries`
* `simpleCUFFT`
* `simpleCUFFT_2d_MGPU`
* `simpleCUFFT_MGPU`
* `simpleCUFFT_callback`
* Updated toolchain for cross-compilation for Tegra QNX platforms.
### CUDA 12.9
* Updated toolchain for cross-compilation for Tegra Linux platforms.
* Added `run_tests.py` utility to exercise all samples. See README.md for details
* Repository has been updated with consistent code formatting across all samples
* Many small code tweaks and bug fixes (see commit history for details)
* Removed the following outdated samples:
* `1_Utilities`
* `bandwidthTest` - this sample was out of date and did not produce accurate results. For bandwidth
testing of NVIDIA GPU platforms, please refer to [NVBandwidth](https://github.com/NVIDIA/nvbandwidth)
### CUDA 12.8
* Updated build system across the repository to CMake. Removed Visual Studio project files and Makefiles.
* Removed the following outdated samples:
* `0_Introduction`
* `c++11_cuda` demonstrating CUDA and C++ 11 interoperability (reason: obsolete)
* `concurrentKernels` demonstrating the ability to run multiple kernels simultaneously (reason: obsolete)
* `cppIntegration` demonstrating calling between .cu and .cpp files (reason: obsolete)
* `cppOverload` demonstrating C++ function overloading (reason: obsolete)
* `simpleSeparateCompilation` demonstrating NVCC compilation to a static library (reason: trivial)
* `simpleTemplates_nvrtc` demonstrating NVRTC usage for `simpleTemplates` sample (reason: redundant)
* `simpleVoteIntrinsics_nvrtc` demonstrating NVRTC usage for `simpleVoteIntrinsics` sample (reason: redundant)
* `2_Concepts_and_Techniques`
* `cuHook` demonstrating dlsym hooks. (reason: incompatible with modern `glibc`)
* `4_CUDA_Libraries`
* `batchedLabelMarkersAndLabelCompressionNPP` demonstrating NPP features (reason: some functionality removed from library)
* `5_Domain_Specific`
* Legacy Direct3D 9 and 10 interoperability samples:
* `fluidsD3D9`
* `simpleD3D10`
* `simpleD3D10RenderTarget`
* `simpleD3D10Texture`
* `simpleD3D9`
* `simpleD3D9Texture`
* `SLID3D10Texture`
* `VFlockingD3D10`
* `8_Platform_Specific/Tegra`
* Temporarily removed the following two samples pending updates:
* `nbody_screen` demonstrating the nbody sample in QNX
* `simpleGLES_screen` demonstrating GLES interop in QNX
* Moved the following Tegra-specific samples to a dedicated subdirectory: `8_Platform_Specific/Tegra`
* `EGLSync_CUDAEvent_Interop`
* `cuDLAErrorReporting`
* `cuDLAHybridMode`
* `cuDLALayerwiseStatsHybrid`
* `cuDLALayerwiseStatsStandalone`
* `cuDLAStandaloneMode`
* `cudaNvSciBufMultiplanar`
* `cudaNvSciNvMedia`
* `fluidsGLES`
* `nbody_opengles`
* `simpleGLES`
* `simpleGLES_EGLOutput`
### CUDA 12.5
### CUDA 12.4
* Added graphConditionalNodes Sample
### CUDA 12.3
* Added cuDLA samples
* Fixed jitLto regression
### CUDA 12.2
* libNVVM samples received updates
* Fixed jitLto Case issues
* Enabled HOST_COMPILER flag to the makefiles for GCC which is untested but may still work.
### CUDA 12.1
* Added new sample for Large Kernels
### CUDA 12.0
* Added new flags for JIT compiling
* Removed deprecated APIs in Hopper Architecture
### CUDA 11.6
* Added new folder structure for samples
* Added support of Visual Studio 2022 to all samples supported on [Windows](#windows-1).
* All CUDA samples are now only available on [GitHub](https://github.com/nvidia/cuda-samples). They are no longer available via CUDA toolkit.
### CUDA 11.5
* Added `cuDLAHybridMode`. Demonstrate usage of cuDLA in hybrid mode.
* Added `cuDLAStandaloneMode`. Demonstrate usage of cuDLA in standalone mode.
* Added `cuDLAErrorReporting`. Demonstrate DLA error detection via CUDA.
* Added `graphMemoryNodes`. Demonstrates memory allocations and frees within CUDA graphs using Graph APIs and Stream Capture APIs.
* Added `graphMemoryFootprint`. Demonstrates how graph memory nodes re-use virtual addresses and physical memory.
* All samples from CUDA toolkit are now available on [GitHub](https://github.com/nvidia/cuda-samples).
### CUDA 11.4 update 1
* Added support for VS Code on linux platform.
### CUDA 11.4
* Added `cdpQuadtree`. Demonstrates Quad Trees implementation using CUDA Dynamic Parallelism.
* Updated `simpleVulkan`, `simpleVulkanMMAP` and `vulkanImageCUDA`. Demonstrates use of SPIR-V shaders.
### CUDA 11.3
* Added `streamOrderedAllocationIPC`. Demonstrates Inter Process Communication using one process per GPU for computation.
* Added `simpleCUBLAS_LU`. Demonstrates batched matrix LU decomposition using cuBLAS API `cublas<t>getrfBatched()`
* Updated `simpleVulkan`. Demonstrates use of timeline semaphore.
* Updated multiple samples to use pinned memory using `cudaMallocHost()`.
### CUDA 11.2
* Added `streamOrderedAllocation`. Demonstrates stream ordered memory allocation on a GPU using cudaMallocAsync and cudaMemPool family of APIs.
* Added `streamOrderedAllocationP2P`. Demonstrates peer-to-peer access of stream ordered memory allocated using cudaMallocAsync and cudaMemPool family of APIs.
* Dropped Visual Studio 2015 support from all the windows supported samples.
* FreeImage is no longer distributed with the CUDA Samples. On Windows, see the [Dependencies](./README.md#freeimage) section for more details on how to set up FreeImage. On Linux, it is recommended to install FreeImage with your distribution's package manager.
* All the samples using CUDA Pipeline & Arrive-wait barriers are been updated to use new `cuda::pipeline` and `cuda::barrier` interfaces.
* Updated all the samples to build with parallel build option `--threads` of `nvcc` cuda compiler.
* Added `cudaNvSciNvMedia`. Demonstrates CUDA-NvMedia interop via NvSciBuf/NvSciSync APIs.
* Added `simpleGL`. Demonstrates interoperability between CUDA and OpenGL.
### CUDA 11.1
* Added `watershedSegmentationNPP`. Demonstrates how to use the NPP watershed segmentation function.
* Added `batchedLabelMarkersAndLabelCompressionNPP`. Demonstrates how to use the NPP label markers generation and label compression functions based on a Union Find (UF) algorithm including both single image and batched image versions.
* Dropped Visual Studio 2012, 2013 support from all the windows supported samples.
* Added kernel performing warp aggregated atomic max in multi buckets using cg::labeled_partition & cg::reduce in `warpAggregatedAtomicsCG`.
* Added extended CG shuffle mechanics to `shfl_scan` sample.
* Added `cudaOpenMP`. Demonstrates how to use OpenMP API to write an application for multiple GPUs.
* Added `simpleZeroCopy`. Demonstrates how to use zero copy, kernels can read and write directly to pinned system memory.
### CUDA 11.0
* Added `dmmaTensorCoreGemm`. Demonstrates double precision GEMM computation using the Double precision Warp Matrix Multiply and Accumulate (WMMA) API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `bf16TensorCoreGemm`. Demonstrates __nv_bfloat16 (e8m7) GEMM computation using the __nv_bfloat16 WMMA API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `tf32TensorCoreGemm`. Demonstrates tf32 (e8m10) GEMM computation using the tf32 WMMA API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `globalToShmemAsyncCopy`. Demonstrates async copy of data from global to shared memory when on compute capability 8.0 or higher. Also demonstrates arrive-wait barrier for synchronization.
* Added `simpleAWBarrier`. Demonstrates arrive wait barriers.
* Added `simpleAttributes`. Demonstrates the stream attributes that affect L2 locality.
* Added warp aggregated atomic multi bucket increments kernel using labeled_partition cooperative groups in `warpAggregatedAtomicsCG` which can be used on compute capability 7.0 and above GPU architectures.
* Added `binaryPartitionCG`. Demonstrates binary partition cooperative groups and reduction within the thread block.
* Added two new reduction kernels in `reduction` one which demonstrates reduce_add_sync intrinstic supported on compute capability 8.0 and another which uses cooperative_groups::reduce function which does thread_block_tile level reduction introduced from CUDA 11.0.
* Added `cudaCompressibleMemory`. Demonstrates compressible memory allocation using cuMemMap API.
* Added `simpleVulkanMMAP`. Demonstrates Vulkan CUDA Interop via cuMemMap APIs.
* Added `concurrentKernels`. Demonstrates the use of CUDA streams for concurrent execution of several kernels on a GPU.
* Dropped Mac OSX support from all samples.
### CUDA 10.2
* Added `simpleD3D11`. Demonstrates CUDA-D3D11 External Resource Interoperability APIs for updating D3D11 buffers from CUDA and synchronization between D3D11 and CUDA with Keyed Mutexes.
* Added `simpleDrvRuntime`. Demonstrates CUDA Driver and Runtime APIs working together to load fatbinary of a CUDA kernel.
* Added `vectorAddMMAP`. Demonstrates how cuMemMap API allows the user to specify the physical properties of their memory while retaining the contiguous nature of their access.
* Added `memMapIPCDrv`. Demonstrates Inter Process Communication using cuMemMap APIs.
* Added `cudaNvSci`. Demonstrates CUDA-NvSciBuf/NvSciSync Interop.
* Added `jacobiCudaGraphs`. Demonstrates Instantiated CUDA Graph Update with Jacobi Iterative Method using different approaches.
* Added `cuSolverSp_LinearSolver`. Demonstrates cuSolverSP's LU, QR and Cholesky factorization.
* Added `MersenneTwisterGP11213`. Demonstrates the Mersenne Twister random number generator GP11213 in cuRAND.
### CUDA 10.1 Update 2
* Added `vulkanImageCUDA`. Demonstrates how to perform Vulkan image - CUDA Interop.
* Added `nvJPEG_encoder`. Demonstrates encoding of jpeg images using NVJPEG Library.
* Added Windows OS support to `nvJPEG` sample.
* Added `boxFilterNPP`. Demonstrates how to use NPP FilterBox function to perform a box filter.
* Added `cannyEdgeDetectorNPP`. Demonstrates the nppiFilterCannyBorder_8u_C1R Canny Edge Detection image filter function.
### CUDA 10.1 Update 1
* Added `NV12toBGRandResize`. Demonstrates how to convert and resize NV12 frames to BGR planars frames using CUDA in batch.
* Added `EGLStream_CUDA_Interop`. Demonstrates data exchange between CUDA and EGL Streams.
* Added `cuSolverDn_LinearSolver`. Demonstrates cuSolverDN's LU, QR and Cholesky factorization.
* Added support of Visual Studio 2019 to all samples supported on [Windows](./README.md#windows-1).
### CUDA 10.1
* Added `immaTensorCoreGemm`. Demonstrates integer GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API for integers employing the Tensor Cores.
* Added `simpleIPC`. Demonstrates Inter Process Communication with one process per GPU for computation.
* Added `nvJPEG`. Demonstrates single and batched decoding of jpeg images using NVJPEG Library.
* Added `bandwidthTest`. It measures the memcopy bandwidth of the GPU and memcpy bandwidth across PCI-e.
* Added `reduction`. Demonstrates several important optimization strategies for Data-Parallel Algorithms like reduction.
* Update all the samples to support CUDA 10.1.
### CUDA 10.0
* Added `simpleCudaGraphs`. Demonstrates CUDA Graphs creation, instantiation and launch using Graphs APIs and Stream Capture APIs.
* Added `conjugateGradientCudaGraphs`. Demonstrates conjugate gradient solver on GPU using CUBLAS and CUSPARSE library calls captured and called using CUDA Graph APIs.
* Added `simpleVulkan`. Demonstrates Vulkan - CUDA Interop.
* Added `simpleD3D12`. Demonstrates DX12 - CUDA Interop.
* Added `UnifiedMemoryPerf`. Demonstrates performance comparision of various memory types involved in system.
* Added `p2pBandwidthLatencyTest`. Demonstrates Peer-To-Peer (P2P) data transfers between pairs of GPUs and computes latency and bandwidth.
* Added `systemWideAtomics`. Demonstrates system wide atomic instructions.
* Added `simpleCUBLASXT`. Demonstrates CUBLAS-XT library which performs GEMM operations over multiple GPUs.
* Added Windows OS support to `conjugateGradientMultiDeviceCG` sample.
* Removed support of Visual Studio 2010 from all samples.
### CUDA 9.2
This is the first release of CUDA Samples on GitHub:
* Added `vectorAdd_nvrtc`. Demonstrates runtime compilation library using NVRTC of a simple vectorAdd kernel.
* Added `warpAggregatedAtomicsCG`. Demonstrates warp aggregated atomics using Cooperative Groups.
* Added `deviceQuery`. Enumerates the properties of the CUDA devices present in the system.
* Added `matrixMul`. Demonstrates a matrix multiplication using shared memory through tiled approach.
* Added `matrixMulDrv`. Demonstrates a matrix multiplication using shared memory through tiled approach, uses CUDA Driver API.
* Added `cudaTensorCoreGemm`. Demonstrates a GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API introduced in CUDA 9, as well as the new Tensor Cores introduced in the Volta chip family.
* Added `simpleVoteIntrinsics` which uses *_sync equivalent of the vote intrinsics _any, _all added since CUDA 9.0.
* Added `shfl_scan` which uses *_sync equivalent of the shfl intrinsics added since CUDA 9.0.
* Added `conjugateGradientMultiBlockCG`. Demonstrates a conjugate gradient solver on GPU using Multi Block Cooperative Groups.
* Added `conjugateGradientMultiDeviceCG`. Demonstrates a conjugate gradient solver on multiple GPUs using Multi Device Cooperative Groups, also uses unified memory prefetching and usage hints APIs.
* Added `simpleCUBLAS`. Demonstrates how perform GEMM operations using CUBLAS library.
* Added `simpleCUFFT`. Demonstrates how perform FFT operations using CUFFT library.

View File

@@ -1,33 +0,0 @@
cmake_minimum_required(VERSION 3.20)
project(cuda-samples LANGUAGES C CXX CUDA)
find_package(CUDAToolkit REQUIRED)
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CUDA_STANDARD 17)
set(CMAKE_CUDA_STANDARD_REQUIRED ON)
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
if(ENABLE_CUDA_DEBUG)
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
else()
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -lineinfo") # add line information to all builds for debug tools (exclusive to -G option)
endif()
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} --extended-lambda")
# Add MSVC-specific flags for standard-conforming preprocessor (required for CCCL)
if(MSVC)
add_compile_options($<$<COMPILE_LANGUAGE:CUDA>:-Xcompiler=/Zc:preprocessor>)
endif()
# Include installation configuration before processing samples
include(cmake/InstallSamples.cmake)
add_subdirectory(cpp)

View File

@@ -1,103 +0,0 @@
# Contributing to the CUDA Samples
Thank you for your interest in contributing to the CUDA Samples!
## Getting Started
1. **Fork & Clone the Repository**:
Fork the reporistory and clone the fork. For more information, check [GitHub's documentation on forking](https://docs.github.com/en/github/getting-started-with-github/fork-a-repo) and [cloning a repository](https://docs.github.com/en/github/creating-cloning-and-archiving-repositories/cloning-a-repository).
## Making Changes
1. **Create a New Branch**:
```bash
git checkout -b your-feature-branch
```
2. **Make Changes**.
3. **Build and Test**:
Ensure changes don't break existing functionality by building and running tests.
For more details on building and testing, refer to the [Building and Testing](#building-and-testing) section below.
4. **Commit Changes**:
```bash
git commit -m "Brief description of the change"
```
## Building and Testing
For information on building a running tests on the samples, please refer to the main [README](README.md)
## Creating a Pull Request
1. Push changes to your fork
2. Create a pull request targeting the `master` branch of the original CUDA Samples repository. Refer to [GitHub's documentation](https://docs.github.com/en/github/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/about-pull-requests) for more information on creating a pull request.
3. Describe the purpose and context of the changes in the pull request description.
## Code Formatting (pre-commit hooks)
The CUDA Samples repository uses [pre-commit](https://pre-commit.com/) to execute all code linters and formatters. These
tools ensure a consistent coding style throughout the project. Using pre-commit ensures that linter
versions and options are aligned for all developers. Additionally, there is a CI check in place to
enforce that committed code follows our standards.
The linters used by the CUDA Samples are listed in `.pre-commit-config.yaml`.
For example, C++ and CUDA code is formatted with [`clang-format`](https://clang.llvm.org/docs/ClangFormat.html).
To use `pre-commit`, install via `conda` or `pip`:
```bash
conda config --add channels conda-forge
conda install pre-commit
```
```bash
pip install pre-commit
```
Then run pre-commit hooks before committing code:
```bash
pre-commit run
```
By default, pre-commit runs on staged files (only changes and additions that will be committed).
To run pre-commit checks on all files, execute:
```bash
pre-commit run --all-files
```
Optionally, you may set up the pre-commit hooks to run automatically when you make a git commit. This can be done by running:
```bash
pre-commit install
```
Now code linters and formatters will be run each time you commit changes.
You can skip these checks with `git commit --no-verify` or with the short version `git commit -n`, althoguh please note
that this may result in pull requests being rejected if subsequent checks fail.
## Review Process
Once submitted, maintainers will be automatically assigned to review the pull request. They might suggest changes or improvements. Constructive feedback is a part of the collaborative process, aimed at ensuring the highest quality code.
For constructive feedback and effective communication during reviews, we recommend following [Conventional Comments](https://conventionalcomments.org/).
Further recommended reading for successful PR reviews:
- [How to Do Code Reviews Like a Human (Part One)](https://mtlynch.io/human-code-reviews-1/)
- [How to Do Code Reviews Like a Human (Part Two)](https://mtlynch.io/human-code-reviews-2/)
## Thank You
Your contributions enhance the CUDA Samples for the entire community. We appreciate your effort and collaboration!

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

File diff suppressed because one or more lines are too long

Binary file not shown.

Binary file not shown.

Binary file not shown.

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

294
Common/dynlink_d3d10.h Normal file
View File

@@ -0,0 +1,294 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//--------------------------------------------------------------------------------------
// File: dynlink_d3d10.h
//
// Shortcut macros and functions for using DX objects
//
// Copyright (c) Microsoft Corporation. All rights reserved
//--------------------------------------------------------------------------------------
#ifndef _DYNLINK_D3D10_H_
#define _DYNLINK_D3D10_H_
// Standard Windows includes
#include <windows.h>
#include <initguid.h>
#include <assert.h>
#include <wchar.h>
#include <mmsystem.h>
#include <commctrl.h> // for InitCommonControls()
#include <shellapi.h> // for ExtractIcon()
#include <new.h> // for placement new
#include <shlobj.h>
#include <math.h>
#include <limits.h>
#include <stdio.h>
// CRT's memory leak detection
#if defined(DEBUG) || defined(_DEBUG)
#include <crtdbg.h>
#endif
// Direct3D9 includes
#include <d3d9.h>
// Direct3D10 includes
#include <dxgi.h>
#include <d3d10_1.h>
#include <d3d10.h>
// XInput includes
#include <xinput.h>
// strsafe.h deprecates old unsecure string functions. If you
// really do not want to it to (not recommended), then uncomment the next line
//#define STRSAFE_NO_DEPRECATE
#ifndef STRSAFE_NO_DEPRECATE
#pragma deprecated("strncpy")
#pragma deprecated("wcsncpy")
#pragma deprecated("_tcsncpy")
#pragma deprecated("wcsncat")
#pragma deprecated("strncat")
#pragma deprecated("_tcsncat")
#endif
#pragma warning( disable : 4996 ) // disable deprecated warning
#include <strsafe.h>
#pragma warning( default : 4996 )
#include <DirectXMath.h>
using namespace DirectX;
//--------------------------------------------------------------------------------------
// Structs
//--------------------------------------------------------------------------------------
struct DXUTD3D9DeviceSettings
{
UINT AdapterOrdinal;
D3DDEVTYPE DeviceType;
D3DFORMAT AdapterFormat;
DWORD BehaviorFlags;
D3DPRESENT_PARAMETERS pp;
};
struct DXUTD3D10DeviceSettings
{
UINT AdapterOrdinal;
D3D10_DRIVER_TYPE DriverType;
UINT Output;
DXGI_SWAP_CHAIN_DESC sd;
UINT32 CreateFlags;
UINT32 SyncInterval;
DWORD PresentFlags;
bool AutoCreateDepthStencil; // DXUT will create the a depth stencil resource and view if true
DXGI_FORMAT AutoDepthStencilFormat;
};
enum DXUTDeviceVersion { DXUT_D3D9_DEVICE, DXUT_D3D10_DEVICE };
struct DXUTDeviceSettings
{
DXUTDeviceVersion ver;
union
{
DXUTD3D9DeviceSettings d3d9; // only valid if ver == DXUT_D3D9_DEVICE
DXUTD3D10DeviceSettings d3d10; // only valid if ver == DXUT_D3D10_DEVICE
};
};
//--------------------------------------------------------------------------------------
// Error codes
//--------------------------------------------------------------------------------------
#define DXUTERR_NODIRECT3D MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0901)
#define DXUTERR_NOCOMPATIBLEDEVICES MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0902)
#define DXUTERR_MEDIANOTFOUND MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0903)
#define DXUTERR_NONZEROREFCOUNT MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0904)
#define DXUTERR_CREATINGDEVICE MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0905)
#define DXUTERR_RESETTINGDEVICE MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0906)
#define DXUTERR_CREATINGDEVICEOBJECTS MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0907)
#define DXUTERR_RESETTINGDEVICEOBJECTS MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x0908)
#define DXUTERR_DEVICEREMOVED MAKE_HRESULT(SEVERITY_ERROR, FACILITY_ITF, 0x090A)
typedef HRESULT(WINAPI *LPCREATEDXGIFACTORY)(REFIID, void **);
typedef HRESULT(WINAPI *LPD3D10CREATEDEVICE)(IDXGIAdapter *, D3D10_DRIVER_TYPE, HMODULE, UINT, UINT32,
ID3D10Device **);
typedef HRESULT(WINAPI *LPD3D10CREATEDEVICE1)(IDXGIAdapter *, D3D10_DRIVER_TYPE, HMODULE, UINT,
D3D10_FEATURE_LEVEL1, UINT, ID3D10Device1 **);
typedef HRESULT(WINAPI *LPD3D10CREATESTATEBLOCK)(ID3D10Device *pDevice, D3D10_STATE_BLOCK_MASK *pStateBlockMask,
ID3D10StateBlock **ppStateBlock);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKUNION)(D3D10_STATE_BLOCK_MASK *pA, D3D10_STATE_BLOCK_MASK *pB,
D3D10_STATE_BLOCK_MASK *pResult);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKINTERSECT)(D3D10_STATE_BLOCK_MASK *pA, D3D10_STATE_BLOCK_MASK *pB,
D3D10_STATE_BLOCK_MASK *pResult);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKDIFFERENCE)(D3D10_STATE_BLOCK_MASK *pA, D3D10_STATE_BLOCK_MASK *pB,
D3D10_STATE_BLOCK_MASK *pResult);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKENABLECAPTURE)(D3D10_STATE_BLOCK_MASK *pMask,
D3D10_DEVICE_STATE_TYPES StateType, UINT RangeStart,
UINT RangeLength);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKDISABLECAPTURE)(D3D10_STATE_BLOCK_MASK *pMask,
D3D10_DEVICE_STATE_TYPES StateType, UINT RangeStart,
UINT RangeLength);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKENABLEALL)(D3D10_STATE_BLOCK_MASK *pMask);
typedef HRESULT(WINAPI *LPD3D10STATEBLOCKMASKDISABLEALL)(D3D10_STATE_BLOCK_MASK *pMask);
typedef BOOL (WINAPI *LPD3D10STATEBLOCKMASKGETSETTING)(D3D10_STATE_BLOCK_MASK *pMask,
D3D10_DEVICE_STATE_TYPES StateType, UINT Entry);
typedef HRESULT(WINAPI *LPD3D10COMPILEEFFECTFROMMEMORY)(void *pData, SIZE_T DataLength, LPCSTR pSrcFileName,
CONST D3D10_SHADER_MACRO *pDefines,
ID3D10Include *pInclude, UINT HLSLFlags, UINT FXFlags,
ID3D10Blob **ppCompiledEffect, ID3D10Blob **ppErrors);
typedef HRESULT(WINAPI *LPD3D10CREATEEFFECTFROMMEMORY)(void *pData, SIZE_T DataLength, UINT FXFlags,
ID3D10Device *pDevice,
ID3D10EffectPool *pEffectPool,
ID3D10Effect **ppEffect);
typedef HRESULT(WINAPI *LPD3D10CREATEEFFECTPOOLFROMMEMORY)(void *pData, SIZE_T DataLength, UINT FXFlags,
ID3D10Device *pDevice, ID3D10EffectPool **ppEffectPool);
typedef HRESULT(WINAPI *LPD3D10CREATEDEVICEANDSWAPCHAIN)(IDXGIAdapter *pAdapter,
D3D10_DRIVER_TYPE DriverType,
HMODULE Software,
UINT Flags,
UINT SDKVersion,
DXGI_SWAP_CHAIN_DESC *pSwapChainDesc,
IDXGISwapChain **ppSwapChain,
ID3D10Device **ppDevice);
typedef HRESULT(WINAPI *LPD3D10CREATEDEVICEANDSWAPCHAIN1)(IDXGIAdapter *pAdapter,
D3D10_DRIVER_TYPE DriverType,
HMODULE Software,
UINT Flags,
D3D10_FEATURE_LEVEL1 HardwareLevel,
UINT SDKVersion,
DXGI_SWAP_CHAIN_DESC *pSwapChainDesc,
IDXGISwapChain **ppSwapChain,
ID3D10Device1 **ppDevice);
// Module and function pointers
static HMODULE g_hModDXGI = NULL;
static HMODULE g_hModD3D10 = NULL;
static HMODULE g_hModD3D101 = NULL;
static LPCREATEDXGIFACTORY sFnPtr_CreateDXGIFactory = NULL;
static LPD3D10CREATESTATEBLOCK sFnPtr_D3D10CreateStateBlock = NULL;
static LPD3D10CREATEDEVICE sFnPtr_D3D10CreateDevice = NULL;
static LPD3D10CREATEDEVICE1 sFnPtr_D3D10CreateDevice1 = NULL;
static LPD3D10STATEBLOCKMASKUNION sFnPtr_D3D10StateBlockMaskUnion = NULL;
static LPD3D10STATEBLOCKMASKINTERSECT sFnPtr_D3D10StateBlockMaskIntersect = NULL;
static LPD3D10STATEBLOCKMASKDIFFERENCE sFnPtr_D3D10StateBlockMaskDifference = NULL;
static LPD3D10STATEBLOCKMASKENABLECAPTURE sFnPtr_D3D10StateBlockMaskEnableCapture = NULL;
static LPD3D10STATEBLOCKMASKDISABLECAPTURE sFnPtr_D3D10StateBlockMaskDisableCapture = NULL;
static LPD3D10STATEBLOCKMASKENABLEALL sFnPtr_D3D10StateBlockMaskEnableAll = NULL;
static LPD3D10STATEBLOCKMASKDISABLEALL sFnPtr_D3D10StateBlockMaskDisableAll = NULL;
static LPD3D10STATEBLOCKMASKGETSETTING sFnPtr_D3D10StateBlockMaskGetSetting = NULL;
static LPD3D10COMPILEEFFECTFROMMEMORY sFnPtr_D3D10CompileEffectFromMemory = NULL;
static LPD3D10CREATEEFFECTFROMMEMORY sFnPtr_D3D10CreateEffectFromMemory = NULL;
static LPD3D10CREATEEFFECTPOOLFROMMEMORY sFnPtr_D3D10CreateEffectPoolFromMemory = NULL;
static LPD3D10CREATEDEVICEANDSWAPCHAIN sFnPtr_D3D10CreateDeviceAndSwapChain = NULL;
static LPD3D10CREATEDEVICEANDSWAPCHAIN1 sFnPtr_D3D10CreateDeviceAndSwapChain1 = NULL;
// unload the D3D10 DLLs
static bool dynlinkUnloadD3D10API(void)
{
if (g_hModD3D10)
{
FreeLibrary(g_hModD3D10);
g_hModD3D10 = NULL;
}
if (g_hModDXGI)
{
FreeLibrary(g_hModDXGI);
g_hModDXGI = NULL;
}
if (g_hModD3D101)
{
FreeLibrary(g_hModD3D101);
g_hModD3D101 = NULL;
}
return true;
}
// Dynamically load the D3D10 DLLs loaded and map the function pointers
static bool dynlinkLoadD3D10API(void)
{
// First check to see if the D3D10 Library is present.
// if it succeeds, then we can call GetProcAddress to grab all of the DX10 functions
g_hModD3D10 = LoadLibrary("d3d10.dll");
if (g_hModD3D10 != NULL)
{
sFnPtr_D3D10CreateStateBlock = (LPD3D10CREATESTATEBLOCK) GetProcAddress(g_hModD3D10, "D3D10CreateStateBlock");
sFnPtr_D3D10CreateDevice = (LPD3D10CREATEDEVICE) GetProcAddress(g_hModD3D10, "D3D10CreateDevice");
sFnPtr_D3D10StateBlockMaskUnion = (LPD3D10STATEBLOCKMASKUNION) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskUnion");
sFnPtr_D3D10StateBlockMaskIntersect = (LPD3D10STATEBLOCKMASKINTERSECT) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskIntersect");
sFnPtr_D3D10StateBlockMaskDifference = (LPD3D10STATEBLOCKMASKDIFFERENCE) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskDifference");
sFnPtr_D3D10StateBlockMaskEnableCapture = (LPD3D10STATEBLOCKMASKENABLECAPTURE) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskEnableCapture");
sFnPtr_D3D10StateBlockMaskDisableCapture = (LPD3D10STATEBLOCKMASKDISABLECAPTURE)GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskDisableCapture");
sFnPtr_D3D10StateBlockMaskEnableAll = (LPD3D10STATEBLOCKMASKENABLEALL) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskEnableAll");
sFnPtr_D3D10StateBlockMaskDisableAll = (LPD3D10STATEBLOCKMASKDISABLEALL) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskDisableAll");
sFnPtr_D3D10StateBlockMaskGetSetting = (LPD3D10STATEBLOCKMASKGETSETTING) GetProcAddress(g_hModD3D10, "D3D10StateBlockMaskGetSetting");
sFnPtr_D3D10CompileEffectFromMemory = (LPD3D10COMPILEEFFECTFROMMEMORY) GetProcAddress(g_hModD3D10, "D3D10CompileEffectFromMemory");
sFnPtr_D3D10CreateEffectFromMemory = (LPD3D10CREATEEFFECTFROMMEMORY) GetProcAddress(g_hModD3D10, "D3D10CreateEffectFromMemory");
sFnPtr_D3D10CreateEffectPoolFromMemory = (LPD3D10CREATEEFFECTPOOLFROMMEMORY) GetProcAddress(g_hModD3D10, "D3D10CreateEffectPoolFromMemory");
sFnPtr_D3D10CreateDeviceAndSwapChain = (LPD3D10CREATEDEVICEANDSWAPCHAIN) GetProcAddress(g_hModD3D10, "D3D10CreateDeviceAndSwapChain");
}
g_hModDXGI = LoadLibrary("dxgi.dll");
if (g_hModDXGI)
{
sFnPtr_CreateDXGIFactory = (LPCREATEDXGIFACTORY) GetProcAddress(g_hModDXGI , "CreateDXGIFactory");
}
// This may fail if this machine isn't Windows Vista SP1 or later
g_hModD3D101 = LoadLibrary("d3d10_1.dll");
if (g_hModD3D101 != NULL)
{
sFnPtr_D3D10CreateDevice1 = (LPD3D10CREATEDEVICE1) GetProcAddress(g_hModD3D101, "D3D10CreateDevice1");
sFnPtr_D3D10CreateDeviceAndSwapChain1 = (LPD3D10CREATEDEVICEANDSWAPCHAIN1) GetProcAddress(g_hModD3D101, "D3D10CreateDeviceAndSwapChain1");
}
if (g_hModD3D10 == NULL || g_hModDXGI == NULL || g_hModD3D101 == NULL)
{
dynlinkUnloadD3D10API();
return false;
}
return true;
}
#endif

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -138,29 +138,26 @@ static const char *_cudaGetErrorEnum(cufftResult error) {
case CUFFT_UNALIGNED_DATA: case CUFFT_UNALIGNED_DATA:
return "CUFFT_UNALIGNED_DATA"; return "CUFFT_UNALIGNED_DATA";
case CUFFT_INCOMPLETE_PARAMETER_LIST:
return "CUFFT_INCOMPLETE_PARAMETER_LIST";
case CUFFT_INVALID_DEVICE: case CUFFT_INVALID_DEVICE:
return "CUFFT_INVALID_DEVICE"; return "CUFFT_INVALID_DEVICE";
case CUFFT_PARSE_ERROR:
return "CUFFT_PARSE_ERROR";
case CUFFT_NO_WORKSPACE: case CUFFT_NO_WORKSPACE:
return "CUFFT_NO_WORKSPACE"; return "CUFFT_NO_WORKSPACE";
case CUFFT_NOT_IMPLEMENTED: case CUFFT_NOT_IMPLEMENTED:
return "CUFFT_NOT_IMPLEMENTED"; return "CUFFT_NOT_IMPLEMENTED";
case CUFFT_LICENSE_ERROR:
return "CUFFT_LICENSE_ERROR";
case CUFFT_NOT_SUPPORTED: case CUFFT_NOT_SUPPORTED:
return "CUFFT_NOT_SUPPORTED"; return "CUFFT_NOT_SUPPORTED";
case CUFFT_MISSING_DEPENDENCY:
return "CUFFT_MISSING_DEPENDENCY";
case CUFFT_NVRTC_FAILURE:
return "CUFFT_NVRTC_FAILURE";
case CUFFT_NVJITLINK_FAILURE:
return "CUFFT_NVJITLINK_FAILURE";
case CUFFT_NVSHMEM_FAILURE:
return "CUFFT_NVSHMEM_FAILURE";
} }
return "<unknown>"; return "<unknown>";
@@ -668,15 +665,6 @@ inline int _ConvertSMVer2Cores(int major, int minor) {
{0x75, 64}, {0x75, 64},
{0x80, 64}, {0x80, 64},
{0x86, 128}, {0x86, 128},
{0x87, 128},
{0x89, 128},
{0x90, 128},
{0xa0, 128},
{0xa1, 128},
{0xa3, 128},
{0xb0, 128},
{0xc0, 128},
{0xc1, 128},
{-1, -1}}; {-1, -1}};
int index = 0; int index = 0;
@@ -723,15 +711,6 @@ inline const char* _ConvertSMVer2ArchName(int major, int minor) {
{0x75, "Turing"}, {0x75, "Turing"},
{0x80, "Ampere"}, {0x80, "Ampere"},
{0x86, "Ampere"}, {0x86, "Ampere"},
{0x87, "Ampere"},
{0x89, "Ada"},
{0x90, "Hopper"},
{0xa0, "Blackwell"},
{0xa1, "Blackwell"},
{0xa3, "Blackwell"},
{0xb0, "Blackwell"},
{0xc0, "Blackwell"},
{0xc1, "Blackwell"},
{-1, "Graphics Device"}}; {-1, "Graphics Device"}};
int index = 0; int index = 0;

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -113,15 +113,6 @@ inline int _ConvertSMVer2CoresDRV(int major, int minor) {
{0x75, 64}, {0x75, 64},
{0x80, 64}, {0x80, 64},
{0x86, 128}, {0x86, 128},
{0x87, 128},
{0x89, 128},
{0x90, 128},
{0xa0, 128},
{0xa1, 128},
{0xa3, 128},
{0xb0, 128},
{0xc0, 128},
{0xc1, 128},
{-1, -1}}; {-1, -1}};
int index = 0; int index = 0;
@@ -244,7 +235,7 @@ inline int gpuGetMaxGflopsDeviceIdDRV() {
} }
unsigned long long compute_perf = unsigned long long compute_perf =
((unsigned long long)multiProcessorCount * sm_per_multiproc * (unsigned long long)(multiProcessorCount * sm_per_multiproc *
clockRate); clockRate);
if (compute_perf > max_compute_perf) { if (compute_perf > max_compute_perf) {
@@ -411,3 +402,4 @@ bool inline findFatbinPath(const char *module_file, std::string &module_path, ch
// end of CUDA Helper Functions // end of CUDA Helper Functions
#endif // COMMON_HELPER_CUDA_DRVAPI_H_ #endif // COMMON_HELPER_CUDA_DRVAPI_H_

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -168,7 +168,7 @@ int waitProcess(Process *process) {
#endif #endif
} }
#if defined(__linux__) || defined(__QNX__) #if defined(__linux__)
int ipcCreateSocket(ipcHandle *&handle, const char *name, int ipcCreateSocket(ipcHandle *&handle, const char *name,
const std::vector<Process> &processes) { const std::vector<Process> &processes) {
int server_fd; int server_fd;
@@ -185,30 +185,25 @@ int ipcCreateSocket(ipcHandle *&handle, const char *name,
return -1; return -1;
} }
char path_name[50]; unlink(name);
bzero(&servaddr, sizeof(servaddr));
// Create unique name for the socket with path if SOCK_FOLDER is set.
sprintf(path_name, "%s/%u", getSocketFolder().c_str(), getpid());
unlink(path_name);
memset(&servaddr, 0, sizeof(servaddr));
servaddr.sun_family = AF_UNIX; servaddr.sun_family = AF_UNIX;
size_t len = strlen(path_name); size_t len = strlen(name);
if (len > (sizeof(servaddr.sun_path) - 1)) { if (len > (sizeof(servaddr.sun_path) - 1)) {
perror("IPC failure: Cannot bind provided name to socket. Name too large"); perror("IPC failure: Cannot bind provided name to socket. Name too large");
return -1; return -1;
} }
strncpy(servaddr.sun_path, path_name, len); strncpy(servaddr.sun_path, name, len);
if (bind(server_fd, (struct sockaddr *)&servaddr, SUN_LEN(&servaddr)) < 0) { if (bind(server_fd, (struct sockaddr *)&servaddr, SUN_LEN(&servaddr)) < 0) {
perror("IPC failure: Binding socket failed"); perror("IPC failure: Binding socket failed");
return -1; return -1;
} }
handle->socketName = new char[strlen(path_name) + 1]; handle->socketName = new char[strlen(name) + 1];
strcpy(handle->socketName, path_name); strcpy(handle->socketName, name);
handle->socket = server_fd; handle->socket = server_fd;
return 0; return 0;
} }
@@ -224,13 +219,13 @@ int ipcOpenSocket(ipcHandle *&handle) {
perror("IPC failure:Socket creation error"); perror("IPC failure:Socket creation error");
return -1; return -1;
} }
memset(&cliaddr, 0, sizeof(cliaddr));
cliaddr.sun_family = AF_UNIX;
char temp[50];
// Create unique name for the socket with path if SOCK_FOLDER is set. bzero(&cliaddr, sizeof(cliaddr));
sprintf(temp, "%s/%u", getSocketFolder().c_str(), getpid()); cliaddr.sun_family = AF_UNIX;
char temp[10];
// Create unique name for the socket.
sprintf(temp, "%u", getpid());
strcpy(cliaddr.sun_path, temp); strcpy(cliaddr.sun_path, temp);
if (bind(sock, (struct sockaddr *)&cliaddr, sizeof(cliaddr)) < 0) { if (bind(sock, (struct sockaddr *)&cliaddr, sizeof(cliaddr)) < 0) {
@@ -267,48 +262,41 @@ int ipcRecvShareableHandle(ipcHandle *handle, ShareableHandle *shHandle) {
// Union to guarantee alignment requirements for control array // Union to guarantee alignment requirements for control array
union { union {
struct cmsghdr cm; struct cmsghdr cm;
// This will not work on QNX as QNX CMSG_SPACE calls __cmsg_alignbytes char control[CMSG_SPACE(sizeof(int))];
// And __cmsg_alignbytes is a runtime function instead of compile-time macros
// char control[CMSG_SPACE(sizeof(int))]
char* control;
} control_un; } control_un;
size_t sizeof_control = CMSG_SPACE(sizeof(int)) * sizeof(char);
control_un.control = (char*) malloc(sizeof_control);
struct cmsghdr *cmptr; struct cmsghdr *cmptr;
ssize_t n; ssize_t n;
int receivedfd; int receivedfd;
char dummy_buffer[1]; char dummy_buffer[1];
ssize_t sendResult; ssize_t sendResult;
msg.msg_control = control_un.control; msg.msg_control = control_un.control;
msg.msg_controllen = sizeof_control; msg.msg_controllen = sizeof(control_un.control);
iov[0].iov_base = (void *)dummy_buffer; iov[0].iov_base = (void *)dummy_buffer;
iov[0].iov_len = sizeof(dummy_buffer); iov[0].iov_len = sizeof(dummy_buffer);
msg.msg_iov = iov; msg.msg_iov = iov;
msg.msg_iovlen = 1; msg.msg_iovlen = 1;
if ((n = recvmsg(handle->socket, &msg, 0)) <= 0) { if ((n = recvmsg(handle->socket, &msg, 0)) <= 0) {
perror("IPC failure: Receiving data over socket failed"); perror("IPC failure: Receiving data over socket failed");
free(control_un.control);
return -1; return -1;
} }
if (((cmptr = CMSG_FIRSTHDR(&msg)) != NULL) && if (((cmptr = CMSG_FIRSTHDR(&msg)) != NULL) &&
(cmptr->cmsg_len == CMSG_LEN(sizeof(int)))) { (cmptr->cmsg_len == CMSG_LEN(sizeof(int)))) {
if ((cmptr->cmsg_level != SOL_SOCKET) || (cmptr->cmsg_type != SCM_RIGHTS)) { if ((cmptr->cmsg_level != SOL_SOCKET) || (cmptr->cmsg_type != SCM_RIGHTS)) {
free(control_un.control);
return -1; return -1;
} }
memmove(&receivedfd, CMSG_DATA(cmptr), sizeof(receivedfd)); memmove(&receivedfd, CMSG_DATA(cmptr), sizeof(receivedfd));
*(int *)shHandle = receivedfd; *(int *)shHandle = receivedfd;
} else { } else {
free(control_un.control);
return -1; return -1;
} }
free(control_un.control);
return 0; return 0;
} }
@@ -331,7 +319,7 @@ int ipcSendDataToServer(ipcHandle *handle, const char *serverName,
ssize_t sendResult; ssize_t sendResult;
struct sockaddr_un serveraddr; struct sockaddr_un serveraddr;
memset(&serveraddr, 0, sizeof(serveraddr)); bzero(&serveraddr, sizeof(serveraddr));
serveraddr.sun_family = AF_UNIX; serveraddr.sun_family = AF_UNIX;
strncpy(serveraddr.sun_path, serverName, sizeof(serveraddr.sun_path) - 1); strncpy(serveraddr.sun_path, serverName, sizeof(serveraddr.sun_path) - 1);
@@ -352,22 +340,19 @@ int ipcSendShareableHandle(ipcHandle *handle,
union { union {
struct cmsghdr cm; struct cmsghdr cm;
char* control; char control[CMSG_SPACE(sizeof(int))];
} control_un; } control_un;
size_t sizeof_control = CMSG_SPACE(sizeof(int)) * sizeof(char);
control_un.control = (char*) malloc(sizeof_control);
struct cmsghdr *cmptr; struct cmsghdr *cmptr;
ssize_t readResult; ssize_t readResult;
struct sockaddr_un cliaddr; struct sockaddr_un cliaddr;
socklen_t len = sizeof(cliaddr); socklen_t len = sizeof(cliaddr);
// Construct client address to send this SHareable handle to // Construct client address to send this SHareable handle to
memset(&cliaddr, 0, sizeof(cliaddr)); bzero(&cliaddr, sizeof(cliaddr));
cliaddr.sun_family = AF_UNIX; cliaddr.sun_family = AF_UNIX;
char temp[20]; char temp[10];
sprintf(temp, "%s/%u", getSocketFolder().c_str(), process); sprintf(temp, "%u", process);
strcpy(cliaddr.sun_path, temp); strcpy(cliaddr.sun_path, temp);
len = sizeof(cliaddr); len = sizeof(cliaddr);
@@ -375,7 +360,7 @@ int ipcSendShareableHandle(ipcHandle *handle,
int sendfd = (int)shareableHandles[data]; int sendfd = (int)shareableHandles[data];
msg.msg_control = control_un.control; msg.msg_control = control_un.control;
msg.msg_controllen = sizeof_control; msg.msg_controllen = sizeof(control_un.control);
cmptr = CMSG_FIRSTHDR(&msg); cmptr = CMSG_FIRSTHDR(&msg);
cmptr->cmsg_len = CMSG_LEN(sizeof(int)); cmptr->cmsg_len = CMSG_LEN(sizeof(int));
@@ -395,11 +380,9 @@ int ipcSendShareableHandle(ipcHandle *handle,
ssize_t sendResult = sendmsg(handle->socket, &msg, 0); ssize_t sendResult = sendmsg(handle->socket, &msg, 0);
if (sendResult <= 0) { if (sendResult <= 0) {
perror("IPC failure: Sending data over socket failed"); perror("IPC failure: Sending data over socket failed");
free(control_un.control);
return -1; return -1;
} }
free(control_un.control);
return 0; return 0;
} }

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -54,27 +54,6 @@
#endif #endif
#include <vector> #include <vector>
// The Unix domain sockets creating folder on QNX has been restricted to qnx6-mounted directories since QNX SDP 8.0.3.
#if defined(__QNX__)
#include <string>
inline std::string getSocketFolder() {
return "/storage";
}
// Simple filesystem compatibility for GCC 8.x
#elif defined(__GNUC__) && __GNUC__ < 9
#include <cstdlib>
#include <string>
inline std::string getSocketFolder() {
const char* tmpdir = std::getenv("TMPDIR");
return tmpdir ? std::string(tmpdir) : "/tmp";
}
#else
#include <filesystem>
inline std::string getSocketFolder() {
return std::filesystem::temp_directory_path().string();
}
#endif
typedef struct sharedMemoryInfo_st { typedef struct sharedMemoryInfo_st {
void *addr; void *addr;
size_t size; size_t size;
@@ -105,7 +84,7 @@ int waitProcess(Process *process);
#define checkIpcErrors(ipcFuncResult) \ #define checkIpcErrors(ipcFuncResult) \
if (ipcFuncResult == -1) { fprintf(stderr, "Failure at %u %s\n", __LINE__, __FILE__); exit(EXIT_FAILURE); } if (ipcFuncResult == -1) { fprintf(stderr, "Failure at %u %s\n", __LINE__, __FILE__); exit(EXIT_FAILURE); }
#if defined(__linux__) || defined(__QNX__) #if defined(__linux__)
struct ipcHandle_st { struct ipcHandle_st {
int socket; int socket;
char *socketName; char *socketName;

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -399,10 +399,10 @@ int getInputDir(std::string &input_dir, const char *executable_path) {
// Search in default paths for input images. // Search in default paths for input images.
std::string pathname = ""; std::string pathname = "";
const char *searchPath[] = { const char *searchPath[] = {
"./images", "./images", "../../../../Samples/<executable_name>/images",
"../../../../cpp/4_CUDA_Libraries/<executable_name>/images", "../../../Samples/<executable_name>/images",
"../../../cpp/4_CUDA_Libraries/<executable_name>/images", "../../Samples/<executable_name>/images"};
"../../cpp/4_CUDA_Libraries/<executable_name>/images"};
for (unsigned int i = 0; i < sizeof(searchPath) / sizeof(char *); ++i) { for (unsigned int i = 0; i < sizeof(searchPath) / sizeof(char *); ++i) {
std::string pathname(searchPath[i]); std::string pathname(searchPath[i]);

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -285,73 +285,14 @@ inline char *sdkFindFilePath(const char *filename,
// the .exe file, a .bat file launching an .exe, a browser .exe launching the // the .exe file, a .bat file launching an .exe, a browser .exe launching the
// .exe or .bat, etc // .exe or .bat, etc
const char *searchPath[] = { const char *searchPath[] = {
"./", // same dir "./", // same dir
"./data/", // same dir "./data/", // same dir
"../../../../Samples/<executable_name>/", // up 4 in tree
"../../../../cpp/<executable_name>/", // up 4 in tree "../../../Samples/<executable_name>/", // up 3 in tree
"../../../cpp/<executable_name>/", // up 3 in tree "../../Samples/<executable_name>/", // up 2 in tree
"../../cpp/<executable_name>/", // up 2 in tree "../../../../Samples/<executable_name>/data/", // up 4 in tree
"../../../Samples/<executable_name>/data/", // up 3 in tree
"../../../../cpp/<executable_name>/data/", // up 4 in tree "../../Samples/<executable_name>/data/", // up 2 in tree
"../../../cpp/<executable_name>/data/", // up 3 in tree
"../../cpp/<executable_name>/data/", // up 2 in tree
"../../../../cpp/0_Introduction/<executable_name>/", // up 4 in tree
"../../../cpp/0_Introduction/<executable_name>/", // up 3 in tree
"../../cpp/0_Introduction/<executable_name>/", // up 2 in tree
"../../../../cpp/1_Utilities/<executable_name>/", // up 4 in tree
"../../../cpp/1_Utilities/<executable_name>/", // up 3 in tree
"../../cpp/1_Utilities/<executable_name>/", // up 2 in tree
"../../../../cpp/2_Concepts_and_Techniques/<executable_name>/", // up 4 in tree
"../../../cpp/2_Concepts_and_Techniques/<executable_name>/", // up 3 in tree
"../../cpp/2_Concepts_and_Techniques/<executable_name>/", // up 2 in tree
"../../../../cpp/3_CUDA_Features/<executable_name>/", // up 4 in tree
"../../../cpp/3_CUDA_Features/<executable_name>/", // up 3 in tree
"../../cpp/3_CUDA_Features/<executable_name>/", // up 2 in tree
"../../../../cpp/4_CUDA_Libraries/<executable_name>/", // up 4 in tree
"../../../cpp/4_CUDA_Libraries/<executable_name>/", // up 3 in tree
"../../cpp/4_CUDA_Libraries/<executable_name>/", // up 2 in tree
"../../../../cpp/5_Domain_Specific/<executable_name>/", // up 4 in tree
"../../../cpp/5_Domain_Specific/<executable_name>/", // up 3 in tree
"../../cpp/5_Domain_Specific/<executable_name>/", // up 2 in tree
"../../../../cpp/6_Performance/<executable_name>/", // up 4 in tree
"../../../cpp/6_Performance/<executable_name>/", // up 3 in tree
"../../cpp/6_Performance/<executable_name>/", // up 2 in tree
"../../../../cpp/0_Introduction/<executable_name>/data/", // up 4 in tree
"../../../cpp/0_Introduction/<executable_name>/data/", // up 3 in tree
"../../cpp/0_Introduction/<executable_name>/data/", // up 2 in tree
"../../../../cpp/1_Utilities/<executable_name>/data/", // up 4 in tree
"../../../cpp/1_Utilities/<executable_name>/data/", // up 3 in tree
"../../cpp/1_Utilities/<executable_name>/data/", // up 2 in tree
"../../../../cpp/2_Concepts_and_Techniques/<executable_name>/data/", // up 4 in tree
"../../../cpp/2_Concepts_and_Techniques/<executable_name>/data/", // up 3 in tree
"../../cpp/2_Concepts_and_Techniques/<executable_name>/data/", // up 2 in tree
"../../../../cpp/3_CUDA_Features/<executable_name>/data/", // up 4 in tree
"../../../cpp/3_CUDA_Features/<executable_name>/data/", // up 3 in tree
"../../cpp/3_CUDA_Features/<executable_name>/data/", // up 2 in tree
"../../../../cpp/4_CUDA_Libraries/<executable_name>/data/", // up 4 in tree
"../../../cpp/4_CUDA_Libraries/<executable_name>/data/", // up 3 in tree
"../../cpp/4_CUDA_Libraries/<executable_name>/data/", // up 2 in tree
"../../../../cpp/5_Domain_Specific/<executable_name>/data/", // up 4 in tree
"../../../cpp/5_Domain_Specific/<executable_name>/data/", // up 3 in tree
"../../cpp/5_Domain_Specific/<executable_name>/data/", // up 2 in tree
"../../../../cpp/6_Performance/<executable_name>/data/", // up 4 in tree
"../../../cpp/6_Performance/<executable_name>/data/", // up 3 in tree
"../../cpp/6_Performance/<executable_name>/data/", // up 2 in tree
"../../../../Common/data/", // up 4 in tree "../../../../Common/data/", // up 4 in tree
"../../../Common/data/", // up 3 in tree "../../../Common/data/", // up 3 in tree
"../../Common/data/" // up 2 in tree "../../Common/data/" // up 2 in tree
@@ -421,7 +362,6 @@ inline char *sdkFindFilePath(const char *filename,
} }
// File not found // File not found
printf("\nerror: sdkFindFilePath: file <%s> not found!\n", filename);
return 0; return 0;
} }

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -258,7 +258,7 @@ namespace nv
s[2] = &r3[0]; s[2] = &r3[0];
s[3] = &r4[0]; s[3] = &r4[0];
int i,j,p,jj; register int i,j,p,jj;
for (i=0; i<4; i++) for (i=0; i<4; i++)
{ {

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -49,11 +49,6 @@
void compileFileToCUBIN(char *filename, int argc, char **argv, char **cubinResult, void compileFileToCUBIN(char *filename, int argc, char **argv, char **cubinResult,
size_t *cubinResultSize, int requiresCGheaders) { size_t *cubinResultSize, int requiresCGheaders) {
if (!filename) {
std::cerr << "\nerror: filename is empty for compileFileToCUBIN()!\n";
exit(1);
}
std::ifstream inputFile(filename, std::ifstream inputFile(filename,
std::ios::in | std::ios::binary | std::ios::ate); std::ios::in | std::ios::binary | std::ios::ate);
@@ -116,12 +111,7 @@ void compileFileToCUBIN(char *filename, int argc, char **argv, char **cubinResul
compileOptions = "--include-path="; compileOptions = "--include-path=";
char *strPath = sdkFindFilePath(HeaderNames, argv[0]); std::string path = sdkFindFilePath(HeaderNames, argv[0]);
if (!strPath) {
std::cerr << "\nerror: header file " << HeaderNames << " not found!\n";
exit(1);
}
std::string path = strPath;
if (!path.empty()) { if (!path.empty()) {
std::size_t found = path.find(HeaderNames); std::size_t found = path.find(HeaderNames);
path.erase(found); path.erase(found);
@@ -130,7 +120,6 @@ void compileFileToCUBIN(char *filename, int argc, char **argv, char **cubinResul
"\nCooperativeGroups headers not found, please install it in %s " "\nCooperativeGroups headers not found, please install it in %s "
"sample directory..\n Exiting..\n", "sample directory..\n Exiting..\n",
argv[0]); argv[0]);
exit(1);
} }
compileOptions += path.c_str(); compileOptions += path.c_str();
compileParams[numCompileOptions] = reinterpret_cast<char *>( compileParams[numCompileOptions] = reinterpret_cast<char *>(
@@ -187,7 +176,6 @@ CUmodule loadCUBIN(char *cubin, int argc, char **argv) {
CUcontext context; CUcontext context;
int major = 0, minor = 0; int major = 0, minor = 0;
char deviceName[256]; char deviceName[256];
CUctxCreateParams ctxCreateParams = {};
// Picks the best CUDA device available // Picks the best CUDA device available
CUdevice cuDevice = findCudaDeviceDRV(argc, (const char **)argv); CUdevice cuDevice = findCudaDeviceDRV(argc, (const char **)argv);
@@ -201,7 +189,7 @@ CUmodule loadCUBIN(char *cubin, int argc, char **argv) {
printf("> GPU Device has SM %d.%d compute capability\n", major, minor); printf("> GPU Device has SM %d.%d compute capability\n", major, minor);
checkCudaErrors(cuInit(0)); checkCudaErrors(cuInit(0));
checkCudaErrors(cuCtxCreate(&context, &ctxCreateParams, 0, cuDevice)); checkCudaErrors(cuCtxCreate(&context, 0, cuDevice));
checkCudaErrors(cuModuleLoadData(&module, cubin)); checkCudaErrors(cuModuleLoadData(&module, cubin));
free(cubin); free(cubin);

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -0,0 +1,128 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
////////////////////////////////////////////////////////////////////////////////
//
// Utility funcs to wrap up saving a surface or the back buffer as a PPM file
// In addition, wraps up a threshold comparision of two PPMs.
//
// These functions are designed to be used to implement an automated QA testing
// for SDK samples.
//
// Author: Bryan Dudash
// Email: sdkfeedback@nvidia.com
//
// Copyright (c) NVIDIA Corporation. All rights reserved.
////////////////////////////////////////////////////////////////////////////////
#include <helper_functions.h>
#include <rendercheck_d3d10.h>
HRESULT CheckRenderD3D10::ActiveRenderTargetToPPM(ID3D10Device *pDevice,
const char *zFileName) {
ID3D10RenderTargetView *pRTV = NULL;
pDevice->OMGetRenderTargets(1, &pRTV, NULL);
ID3D10Resource *pSourceResource = NULL;
pRTV->GetResource(&pSourceResource);
return ResourceToPPM(pDevice, pSourceResource, zFileName);
}
HRESULT CheckRenderD3D10::ResourceToPPM(ID3D10Device *pDevice,
ID3D10Resource *pResource,
const char *zFileName) {
D3D10_RESOURCE_DIMENSION rType;
pResource->GetType(&rType);
if (rType != D3D10_RESOURCE_DIMENSION_TEXTURE2D) {
printf("SurfaceToPPM: pResource is not a 2D texture! Aborting...\n");
return E_FAIL;
}
ID3D10Texture2D *pSourceTexture = (ID3D10Texture2D *)pResource;
ID3D10Texture2D *pTargetTexture = NULL;
D3D10_TEXTURE2D_DESC desc;
pSourceTexture->GetDesc(&desc);
desc.BindFlags = 0;
desc.CPUAccessFlags = D3D10_CPU_ACCESS_READ;
desc.Usage = D3D10_USAGE_STAGING;
if (FAILED(pDevice->CreateTexture2D(&desc, NULL, &pTargetTexture))) {
printf(
"SurfaceToPPM: Unable to create target Texture resoruce! Aborting... "
"\n");
return E_FAIL;
}
pDevice->CopyResource(pTargetTexture, pSourceTexture);
D3D10_MAPPED_TEXTURE2D mappedTex2D;
pTargetTexture->Map(0, D3D10_MAP_READ, 0, &mappedTex2D);
// Need to convert from dx pitch to pitch=width
unsigned char *pPPMData = new unsigned char[desc.Width * desc.Height * 4];
for (unsigned int iHeight = 0; iHeight < desc.Height; iHeight++) {
memcpy(
&(pPPMData[iHeight * desc.Width * 4]),
(unsigned char *)(mappedTex2D.pData) + iHeight * mappedTex2D.RowPitch,
desc.Width * 4);
}
pTargetTexture->Unmap(0);
// Prepends the PPM header info and bumps byte data afterwards
sdkSavePPM4ub(zFileName, pPPMData, desc.Width, desc.Height);
delete[] pPPMData;
pTargetTexture->Release();
return S_OK;
}
bool CheckRenderD3D10::PPMvsPPM(const char *src_file, const char *ref_file,
const char *exec_path, const float epsilon,
const float threshold) {
char *ref_file_path = sdkFindFilePath(ref_file, exec_path);
if (ref_file_path == NULL) {
printf(
"CheckRenderD3D10::PPMvsPPM unable to find <%s> in <%s> Aborting "
"comparison!\n",
ref_file, exec_path);
printf(">>> Check info.xml and [project//data] folder <%s> <<<\n",
ref_file);
printf("Aborting comparison!\n");
printf(" FAILURE!\n");
return false;
}
return (sdkComparePPM(src_file, ref_file_path, epsilon, threshold, true) ==
true);
}

View File

@@ -0,0 +1,53 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#pragma once
#ifndef _RENDERCHECK_D3D10_H_
#define _RENDERCHECK_D3D10_H_
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <d3d10.h>
class CheckRenderD3D10 {
public:
CheckRenderD3D10() {}
static HRESULT ActiveRenderTargetToPPM(ID3D10Device *pDevice,
const char *zFileName);
static HRESULT ResourceToPPM(ID3D10Device *pDevice, ID3D10Resource *pResource,
const char *zFileName);
static bool PPMvsPPM(const char *src_file, const char *ref_file,
const char *exec_path, const float epsilon,
const float threshold = 0.0f);
};
#endif

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

167
Common/rendercheck_d3d9.cpp Normal file
View File

@@ -0,0 +1,167 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
////////////////////////////////////////////////////////////////////////////////
//
// Utility funcs to wrap up savings a surface or the back buffer as a PPM file
// In addition, wraps up a threshold comparision of two PPMs.
//
// These functions are designed to be used to implement an automated QA testing
// for SDK samples.
//
// Author: Bryan Dudash
// Email: sdkfeedback@nvidia.com
//
// Copyright (c) NVIDIA Corporation. All rights reserved.
////////////////////////////////////////////////////////////////////////////////
#include <helper_functions.h>
#include <rendercheck_d3d9.h>
// originally copied from checkrender_gl.cpp and slightly modified
bool CheckRenderD3D9::PPMvsPPM(const char *src_file, const char *ref_file,
const char *exec_path, const float epsilon,
const float threshold) {
char *ref_file_path = sdkFindFilePath(ref_file, exec_path);
if (ref_file_path == NULL) {
printf(
"CheckRenderD3D9::PPMvsPPM unable to find <%s> in <%s> Aborting "
"comparison!\n",
ref_file, exec_path);
printf(">>> Check info.xml and [project//data] folder <%s> <<<\n",
ref_file);
printf("Aborting comparison!\n");
printf(" FAILURE!\n");
return false;
}
return (sdkComparePPM(src_file, ref_file_path, epsilon, threshold, true) ==
true);
};
HRESULT CheckRenderD3D9::BackbufferToPPM(IDirect3DDevice9 *pDevice,
const char *zFileName) {
IDirect3DSurface9 *pSurface = NULL;
if (FAILED(
pDevice->GetBackBuffer(0, 0, D3DBACKBUFFER_TYPE_MONO, &pSurface))) {
printf("Unable to get the back buffer. Aborting...\n");
return E_FAIL;
}
// D3DXSaveSurfaceToFile("C:\\bing.dds",D3DXIFF_DDS,pSurface,NULL,NULL);
HRESULT hr = S_OK;
hr = SurfaceToPPM(pDevice, pSurface, zFileName);
pSurface->Release();
return hr;
}
HRESULT CheckRenderD3D9::SurfaceToPPM(IDirect3DDevice9 *pDevice,
IDirect3DSurface9 *pSurface,
const char *zFileName) {
D3DSURFACE_DESC pDesc;
pSurface->GetDesc(&pDesc);
// $$ For now only support common 8bit formats. TODO: support for more
// complex formats via conversion?
if (!(pDesc.Format == D3DFMT_A8R8G8B8 || pDesc.Format == D3DFMT_X8R8G8B8)) {
return E_INVALIDARG;
}
IDirect3DTexture9 *pTargetTex = NULL;
if (FAILED(pDevice->CreateTexture(pDesc.Width, pDesc.Height, 1,
D3DUSAGE_DYNAMIC, pDesc.Format,
D3DPOOL_SYSTEMMEM, &pTargetTex, NULL))) {
printf("Unable to create texture for surface transfer! Aborting...\n");
return E_FAIL;
}
IDirect3DSurface9 *pTargetSurface = NULL;
if (FAILED(pTargetTex->GetSurfaceLevel(0, &pTargetSurface))) {
printf("Unable to get surface for surface transfer! Aborting...\n");
return E_FAIL;
}
// This is required because we cannot lock a D3DPOOL_DEAULT surface directly.
// So, we copy to our sysmem surface.
if (FAILED(pDevice->GetRenderTargetData(pSurface, pTargetSurface))) {
printf(
"Unable to GetRenderTargetData() for surface transfer! Aborting...\n");
return E_FAIL;
}
D3DLOCKED_RECT lockedRect;
HRESULT hr = pTargetSurface->LockRect(&lockedRect, NULL, 0);
// Need to convert from dx pitch to pitch=width
//
// $ PPM is BGR and not RGB it seems. Saved image looks "funny" in viewer(red
// and blue swapped), but since ref will be dumped using same method, this is
// ok.
// however, if we want the saved image to be properly colored, then we
// can swizzle the color bytes here.
unsigned char *pPPMData = new unsigned char[pDesc.Width * pDesc.Height * 4];
for (unsigned int iHeight = 0; iHeight < pDesc.Height; iHeight++) {
#if 1 // swizzle to implment RGB to BGR conversion.
for (unsigned int iWidth = 0; iWidth < pDesc.Width; iWidth++) {
DWORD color = *(DWORD *)((unsigned char *)(lockedRect.pBits) +
iHeight * lockedRect.Pitch + iWidth * 4);
// R<->B, [7:0] <-> [23:16], swizzle
color = ((color & 0xFF) << 16) | (color & 0xFF00) |
((color & 0xFF0000) >> 16) | (color & 0xFF000000);
memcpy(&(pPPMData[(iHeight * pDesc.Width + iWidth) * 4]),
(unsigned char *)&color, 4);
}
#else
memcpy(&(pPPMData[iHeight * pDesc.Width * 4]),
(unsigned char *)(lockedRect.pBits) + iHeight * lockedRect.Pitch,
pDesc.Width * 4);
#endif
}
pTargetSurface->UnlockRect();
// Prepends the PPM header info and bumps byte data afterwards
sdkSavePPM4ub(zFileName, pPPMData, pDesc.Width, pDesc.Height);
delete[] pPPMData;
pTargetSurface->Release();
pTargetTex->Release();
return S_OK;
}

54
Common/rendercheck_d3d9.h Normal file
View File

@@ -0,0 +1,54 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#pragma once
#ifndef _RENDERCHECK_D3D9_H_
#define _RENDERCHECK_D3D9_H_
#include <assert.h>
#include <d3d9.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
class CheckRenderD3D9 {
public:
CheckRenderD3D9() {}
static HRESULT BackbufferToPPM(IDirect3DDevice9 *pDevice,
const char *zFileName);
static HRESULT SurfaceToPPM(IDirect3DDevice9 *pDevice,
IDirect3DSurface9 *pSurface,
const char *zFileName);
static bool PPMvsPPM(const char *src_file, const char *ref_file,
const char *exec_path, const float epsilon,
const float threshold = 0.0f);
};
#endif

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

File diff suppressed because it is too large Load Diff

View File

@@ -1,4 +1,4 @@
Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
Redistribution and use in source and binary forms, with or without Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions modification, are permitted provided that the following conditions

69
Makefile Normal file
View File

@@ -0,0 +1,69 @@
###############################################################################
#
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
###############################################################################
#
# CUDA Samples
#
###############################################################################
TARGET_ARCH ?= $(shell uname -m)
# Project folders that contain CUDA samples
PROJECTS ?= $(shell find Samples -name Makefile)
FILTER_OUT :=
PROJECTS := $(filter-out $(FILTER_OUT),$(PROJECTS))
%.ph_build :
+@$(MAKE) -C $(dir $*) $(MAKECMDGOALS)
%.ph_clean :
+@$(MAKE) -C $(dir $*) clean $(USE_DEVICE)
%.ph_clobber :
+@$(MAKE) -C $(dir $*) clobber $(USE_DEVICE)
%.ph_run :
+@$(MAKE) -C $(dir $*) run
all: $(addsuffix .ph_build,$(PROJECTS))
@echo "Finished building CUDA samples"
build: $(addsuffix .ph_build,$(PROJECTS))
tidy:
@find * | egrep "#" | xargs rm -f
@find * | egrep "\~" | xargs rm -f
clean: tidy $(addsuffix .ph_clean,$(PROJECTS))
clobber: clean $(addsuffix .ph_clobber,$(PROJECTS))
run: $(addsuffix .ph_run,$(PROJECTS))

716
README.md
View File

@@ -1,20 +1,130 @@
# CUDA Samples # CUDA Samples
Samples for CUDA Developers which demonstrates features in CUDA Toolkit. This version supports [CUDA Toolkit 13.2](https://developer.nvidia.com/cuda-downloads). Samples for CUDA Developers which demonstrates features in CUDA Toolkit. This version supports [CUDA Toolkit 11.5](https://developer.nvidia.com/cuda-downloads).
## Release Notes ## Release Notes
This section describes the release notes for the CUDA Samples on GitHub only. This section describes the release notes for the CUDA Samples on GitHub only.
### Change Log ### CUDA 11.5
* Added `cuDLAHybridMode`. Demonstrate usage of cuDLA in hybrid mode.
* Added `cuDLAStandaloneMode`. Demonstrate usage of cuDLA in standalone mode.
* Added `cuDLAErrorReporting`. Demonstrate DLA error detection via CUDA.
* Added `graphMemoryNodes`. Demonstrates memory allocations and frees within CUDA graphs using Graph APIs and Stream Capture APIs.
* Added `graphMemoryFootprint`. Demonstrates how graph memory nodes re-use virtual addresses and physical memory.
* All samples from CUDA toolkit are now available on [GitHub](https://github.com/nvidia/cuda-samples).
### [Revision History](./CHANGELOG.md) ### CUDA 11.4 update 1
* Added support for VS Code on linux platform.
### CUDA 11.4
* Added `cdpQuadtree`. Demonstrates Quad Trees implementation using CUDA Dynamic Parallelism.
* Updated `simpleVulkan`, `simpleVulkanMMAP` and `vulkanImageCUDA`. Demonstrates use of SPIR-V shaders.
### CUDA 11.3
* Added `streamOrderedAllocationIPC`. Demonstrates Inter Process Communication using one process per GPU for computation.
* Added `simpleCUBLAS_LU`. Demonstrates batched matrix LU decomposition using cuBLAS API `cublas<t>getrfBatched()`
* Updated `simpleVulkan`. Demonstrates use of timeline semaphore.
* Updated multiple samples to use pinned memory using `cudaMallocHost()`.
### CUDA 11.2
* Added `streamOrderedAllocation`. Demonstrates stream ordered memory allocation on a GPU using cudaMallocAsync and cudaMemPool family of APIs.
* Added `streamOrderedAllocationP2P`. Demonstrates peer-to-peer access of stream ordered memory allocated using cudaMallocAsync and cudaMemPool family of APIs.
* Dropped Visual Studio 2015 support from all the windows supported samples.
* FreeImage is no longer distributed with the CUDA Samples. On Windows, see the [Dependencies](#freeimage) section for more details on how to set up FreeImage. On Linux, it is recommended to install FreeImage with your distribution's package manager.
* All the samples using CUDA Pipeline & Arrive-wait barriers are been updated to use new `cuda::pipeline` and `cuda::barrier` interfaces.
* Updated all the samples to build with parallel build option `--threads` of `nvcc` cuda compiler.
* Added `cudaNvSciNvMedia`. Demonstrates CUDA-NvMedia interop via NvSciBuf/NvSciSync APIs.
* Added `simpleGL`. Demonstrates interoperability between CUDA and OpenGL.
### CUDA 11.1
* Added `watershedSegmentationNPP`. Demonstrates how to use the NPP watershed segmentation function.
* Added `batchedLabelMarkersAndLabelCompressionNPP`. Demonstrates how to use the NPP label markers generation and label compression functions based on a Union Find (UF) algorithm including both single image and batched image versions.
* Dropped Visual Studio 2012, 2013 support from all the windows supported samples.
* Added kernel performing warp aggregated atomic max in multi buckets using cg::labeled_partition & cg::reduce in `warpAggregatedAtomicsCG`.
* Added extended CG shuffle mechanics to `shfl_scan` sample.
* Added `cudaOpenMP`. Demonstrates how to use OpenMP API to write an application for multiple GPUs.
* Added `simpleZeroCopy`. Demonstrates how to use zero copy, kernels can read and write directly to pinned system memory.
### CUDA 11.0
* Added `dmmaTensorCoreGemm`. Demonstrates double precision GEMM computation using the Double precision Warp Matrix Multiply and Accumulate (WMMA) API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `bf16TensorCoreGemm`. Demonstrates __nv_bfloat16 (e8m7) GEMM computation using the __nv_bfloat16 WMMA API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `tf32TensorCoreGemm`. Demonstrates tf32 (e8m10) GEMM computation using the tf32 WMMA API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `globalToShmemAsyncCopy`. Demonstrates async copy of data from global to shared memory when on compute capability 8.0 or higher. Also demonstrates arrive-wait barrier for synchronization.
* Added `simpleAWBarrier`. Demonstrates arrive wait barriers.
* Added `simpleAttributes`. Demonstrates the stream attributes that affect L2 locality.
* Added warp aggregated atomic multi bucket increments kernel using labeled_partition cooperative groups in `warpAggregatedAtomicsCG` which can be used on compute capability 7.0 and above GPU architectures.
* Added `binaryPartitionCG`. Demonstrates binary partition cooperative groups and reduction within the thread block.
* Added two new reduction kernels in `reduction` one which demonstrates reduce_add_sync intrinstic supported on compute capability 8.0 and another which uses cooperative_groups::reduce function which does thread_block_tile level reduction introduced from CUDA 11.0.
* Added `cudaCompressibleMemory`. Demonstrates compressible memory allocation using cuMemMap API.
* Added `simpleVulkanMMAP`. Demonstrates Vulkan CUDA Interop via cuMemMap APIs.
* Added `concurrentKernels`. Demonstrates the use of CUDA streams for concurrent execution of several kernels on a GPU.
* Dropped Mac OSX support from all samples.
### CUDA 10.2
* Added `simpleD3D11`. Demonstrates CUDA-D3D11 External Resource Interoperability APIs for updating D3D11 buffers from CUDA and synchronization between D3D11 and CUDA with Keyed Mutexes.
* Added `simpleDrvRuntime`. Demonstrates CUDA Driver and Runtime APIs working together to load fatbinary of a CUDA kernel.
* Added `vectorAddMMAP`. Demonstrates how cuMemMap API allows the user to specify the physical properties of their memory while retaining the contiguous nature of their access.
* Added `memMapIPCDrv`. Demonstrates Inter Process Communication using cuMemMap APIs.
* Added `cudaNvSci`. Demonstrates CUDA-NvSciBuf/NvSciSync Interop.
* Added `jacobiCudaGraphs`. Demonstrates Instantiated CUDA Graph Update with Jacobi Iterative Method using different approaches.
* Added `cuSolverSp_LinearSolver`. Demonstrates cuSolverSP's LU, QR and Cholesky factorization.
* Added `MersenneTwisterGP11213`. Demonstrates the Mersenne Twister random number generator GP11213 in cuRAND.
### CUDA 10.1 Update 2
* Added `vulkanImageCUDA`. Demonstrates how to perform Vulkan image - CUDA Interop.
* Added `nvJPEG_encoder`. Demonstrates encoding of jpeg images using NVJPEG Library.
* Added Windows OS support to `nvJPEG` sample.
* Added `boxFilterNPP`. Demonstrates how to use NPP FilterBox function to perform a box filter.
* Added `cannyEdgeDetectorNPP`. Demonstrates the nppiFilterCannyBorder_8u_C1R Canny Edge Detection image filter function.
### CUDA 10.1 Update 1
* Added `NV12toBGRandResize`. Demonstrates how to convert and resize NV12 frames to BGR planars frames using CUDA in batch.
* Added `EGLStream_CUDA_Interop`. Demonstrates data exchange between CUDA and EGL Streams.
* Added `cuSolverDn_LinearSolver`. Demonstrates cuSolverDN's LU, QR and Cholesky factorization.
* Added support of Visual Studio 2019 to all samples supported on [Windows](#windows-1).
### CUDA 10.1
* Added `immaTensorCoreGemm`. Demonstrates integer GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API for integers employing the Tensor Cores.
* Added `simpleIPC`. Demonstrates Inter Process Communication with one process per GPU for computation.
* Added `nvJPEG`. Demonstrates single and batched decoding of jpeg images using NVJPEG Library.
* Added `bandwidthTest`. It measures the memcopy bandwidth of the GPU and memcpy bandwidth across PCI-e.
* Added `reduction`. Demonstrates several important optimization strategies for Data-Parallel Algorithms like reduction.
* Update all the samples to support CUDA 10.1.
### CUDA 10.0
* Added `simpleCudaGraphs`. Demonstrates CUDA Graphs creation, instantiation and launch using Graphs APIs and Stream Capture APIs.
* Added `conjugateGradientCudaGraphs`. Demonstrates conjugate gradient solver on GPU using CUBLAS and CUSPARSE library calls captured and called using CUDA Graph APIs.
* Added `simpleVulkan`. Demonstrates Vulkan - CUDA Interop.
* Added `simpleD3D12`. Demonstrates DX12 - CUDA Interop.
* Added `UnifiedMemoryPerf`. Demonstrates performance comparision of various memory types involved in system.
* Added `p2pBandwidthLatencyTest`. Demonstrates Peer-To-Peer (P2P) data transfers between pairs of GPUs and computes latency and bandwidth.
* Added `systemWideAtomics`. Demonstrates system wide atomic instructions.
* Added `simpleCUBLASXT`. Demonstrates CUBLAS-XT library which performs GEMM operations over multiple GPUs.
* Added Windows OS support to `conjugateGradientMultiDeviceCG` sample.
* Removed support of Visual Studio 2010 from all samples.
### CUDA 9.2
This is the first release of CUDA Samples on GitHub:
* Added `vectorAdd_nvrtc`. Demonstrates runtime compilation library using NVRTC of a simple vectorAdd kernel.
* Added `warpAggregatedAtomicsCG`. Demonstrates warp aggregated atomics using Cooperative Groups.
* Added `deviceQuery`. Enumerates the properties of the CUDA devices present in the system.
* Added `matrixMul`. Demonstrates a matrix multiplication using shared memory through tiled approach.
* Added `matrixMulDrv`. Demonstrates a matrix multiplication using shared memory through tiled approach, uses CUDA Driver API.
* Added `cudaTensorCoreGemm`. Demonstrates a GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API introduced in CUDA 9, as well as the new Tensor Cores introduced in the Volta chip family.
* Added `simpleVoteIntrinsics` which uses *_sync equivalent of the vote intrinsics _any, _all added since CUDA 9.0.
* Added `shfl_scan` which uses *_sync equivalent of the shfl intrinsics added since CUDA 9.0.
* Added `conjugateGradientMultiBlockCG`. Demonstrates a conjugate gradient solver on GPU using Multi Block Cooperative Groups.
* Added `conjugateGradientMultiDeviceCG`. Demonstrates a conjugate gradient solver on multiple GPUs using Multi Device Cooperative Groups, also uses unified memory prefetching and usage hints APIs.
* Added `simpleCUBLAS`. Demonstrates how perform GEMM operations using CUBLAS library.
* Added `simpleCUFFT`. Demonstrates how perform FFT operations using CUFFT library.
## Getting Started ## Getting Started
### Prerequisites ### Prerequisites
Download and install the [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads) for your corresponding platform. Download and install the [CUDA Toolkit 11.5](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
For system requirements and installation instructions of cuda toolkit, please refer to the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/), and the [Windows Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html). For system requirements and installation instructions of cuda toolkit, please refer to the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/), and the [Windows Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html).
### Getting the CUDA Samples ### Getting the CUDA Samples
@@ -28,453 +138,147 @@ Without using git the easiest way to use these samples is to download the zip fi
## Building CUDA Samples ## Building CUDA Samples
### Building CUDA Samples
The CUDA Samples are built using CMake. Follow the instructions below for building on Linux, Windows, and for cross-compilation to Tegra devices.
### Linux
Ensure that CMake (version 3.20 or later) is installed. Install it using your package manager if necessary:
e.g.
```sudo apt install cmake```
Navigate to the root of the cloned repository and create a build directory:
```
mkdir build && cd build
```
Configure the project with CMake:
```
cmake ..
```
Build the samples:
```
make -j$(nproc)
```
Run the samples from their respective directories in the build folder. You can also follow this process from and subdirectory of the samples repo, or from within any individual sample.
### Windows ### Windows
Language services for CMake are available in Visual Studio 2019 version 16.5 or later, and you can directly import the CUDA samples repository from either the root level or from any The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format:
subdirectory or individual sample.
To build from the command line, open the `x64 Native Tools Command Prompt for VS` provided with your Visual Studio installation.
Navigate to the root of the cloned repository and create a build directory:
``` ```
mkdir build && cd build *_vs<version>.sln - for Visual Studio <version>
``` ```
Configure the project with CMake - for example: Complete samples solution files exist at parent directory of the repo:
Each individual sample has its own set of solution files at:
`<CUDA_SAMPLES_REPO>\Samples\<sample_dir>\`
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
``` ```
cmake .. -G "Visual Studio 16 2019" -A x64 $ cd <sample_dir>
$ make
``` ```
Open the generated solution file CUDA_Samples.sln in Visual Studio. Build the samples by selecting the desired configuration (e.g., Debug or Release) and pressing F7 (Build Solution). The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, ppc64le, armv7l, aarch64.
Run the samples from the output directories specified in Visual Studio. By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=ppc64le` <br/> `$ make TARGET_ARCH=armv7l` <br/> `$ make TARGET_ARCH=aarch64` <br/>
### Enabling On-GPU Debugging See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details on cross platform compilation of cuda samples.
* **dbg=1** - build with debug symbols
NVIDIA GPUs support on-GPU debugging through cuda-gdb. Enabling this may significantly affect application performance as certain compiler optimizations are disabled ```
in this configuration, hence it's not on by default. Enablement of on-device debugging is controlled via the `-G` switch to nvcc. $ make dbg=1
```
To enable cuda-gdb for samples builds, define the `ENABLE_CUDA_DEBUG` flag on the CMake command line. For example: * **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
``` $ make SMS="50 60"
cmake -DENABLE_CUDA_DEBUG=True ... ```
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
### Platform-Specific Samples ```
$ make HOST_COMPILER=g++
Some CUDA samples are specific to certain platforms, and require passing flags into CMake to enable. In particular, we define the following platform-specific flags: ```
* `BUILD_TEGRA` - for Tegra-specific samples
To build these samples, set the variables either on the command line or through your CMake GUI. For example:
```
cmake -DBUILD_TEGRA=True ..
```
### Cross-Compilation for Tegra Platforms
Install the NVIDIA toolchain and cross-compilation environment for Tegra devices as described in the Tegra Development Guide.
Ensure that CMake (version 3.20 or later) is installed.
Navigate to the root of the cloned repository and create a build directory:
```
mkdir build && cd build
```
Configure the project with CMake, specifying the Tegra toolchain file. And you can use -DTARGET_FS to point to the target file system root path for necessary include and library files:
```
cmake .. -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/toolchain-aarch64-linux.cmake -DTARGET_FS=/path/to/target/system/file/system
```
Build the samples:
```
make -j$(nproc)
```
Transfer the built binaries to the Tegra device and execute them there.
### Cross Building for Automotive Linux Platforms from the DriveOS Docker containers
To build CUDA samples to the target platform from the DriveOS Docker containers, use the following instructions.
Mount the target Root Filesystem (RFS) in the container so that the CUDA cmake process has the correct paths to CUDA and other system libraries required to build the samples.
Create a temporary directory, `<temp>` is any temporary directory of your choosing, for example, you can use `/drive/temp`:
```
$ mkdir /drive/<temp>
```
Mount the filesystem by running the following command:
```
$ mount /drive/drive-linux/filesystem/targetfs-images/dev_nsr_desktop_ubuntu-24.04_thor_rfs.img /drive/temp
```
Configure the project by running the following cmake command:
```
$ mkdir build && cd build
$ cmake .. -DBUILD_TEGRA=True \
-DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc \
-DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/toolchain-aarch64-linux.cmake \
-DTARGET_FS=/drive/temp \
-DCMAKE_LIBRARY_PATH=/drive/temp/usr/local/cuda-13.1/thor/lib64/ \
-DCMAKE_INCLUDE_PATH=/drive/temp/usr/local/cuda-13.1/thor/include/
```
Please note that the following libraries are not pre-installed in the DriveOS dev-nsr target filesystem:
* libdrm-dev
* Vulkan
This causes the cmake command to throw errors related to the missing files, and as a result, the related samples will not build in later steps. This issue will be addressed in a future DriveOS release.
To build the samples with ignore the error mentioned above, you can use `--ignore-errors`/`--keep-going` or comment out the comment out the corresponding `add_subdirectory` command in the CMakeLists.txt in the parent folder for the samples requiring Vulkan and libdrm_dev:
```
$ make -j$(nproc) --ignore-errors # or --keep-going
```
```
# In cpp/5_Domain_Specific/CMakeList.txt
# add_subdirectory(simpleGL)
# add_subdirectory(simpleVulkan)
# add_subdirectory(simpleVulkanMMAP)
# In cpp/8_Platform_Specific/Tegra/CMakeList.txt
# add_subdirectory(simpleGLES_EGLOutput)
```
### QNX
Cross-compilation for QNX with CMake is supported in the CUDA 13.0 samples release and newer. An example build for
the Tegra Thor QNX platform might look like this:
```
$ mkdir build
$ cd build
QNX_HOST=/path/to/qnx/host \
QNX_TARGET=/path/to/qnx/target \
cmake .. \
-DBUILD_TEGRA=True \
-DCMAKE_CUDA_COMPILER=/usr/local/cuda-safe-13.0/bin/nvcc \
-DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/toolchain-aarch64-qnx.cmake \
-DCMAKE_LIBRARY_PATH=/usr/local/cuda-safe-13.0/thor/targets/aarch64-qnx/lib/stubs/ \
-DCMAKE_INCLUDE_PATH=/usr/local/cuda-safe-13.0/thor/targets/aarch64-qnx/include/
```
### Forward Compatibility
To build samples with new CUDA Toolkit(CUDA 13.0 or later) and UMD(Version 580 or later) and old KMD(Version 550 or earlier),you need to set the `CMAKE_PREFIX_PATH` for using new driver library, the command might like this:
```
cmake -DCMAKE_PREFIX_PATH=/usr/local/cuda/lib64/stubs/ ..
```
## CUDA Python samples
The repository includes **Python** examples under the [`python/`](./python) directory. **These samples are cuda.core–focused:** they use [CUDA Python](https://nvidia.github.io/cuda-python/), with [`cuda.core`](https://nvidia.github.io/cuda-python/cuda-core/latest/) for devices, programs, launches, and memory, alongside NumPy, CuPy, or framework interop where each sample notes.
**Layout (same themes as the C++ samples):**
| Directory | Contents |
|-----------|----------|
| `python/1_GettingStarted/` | Introductory scripts (e.g. `vectorAdd`, `deviceQuery`, `systemInfo`, image blur with unified memory, NumPy vs CuPy). |
| `python/2_CoreConcepts/` | Algorithms and techniques (e.g. reductions, histograms, FFT, stream overlap, `memoryResources`, `cudaGraphs`, `jitLtoLinking`, `tmaTensorMap`). |
| `python/3_FrameworkInterop/` | Integration with PyTorch and TensorFlow. |
| `python/4_DistributedComputing/` | Multi-GPU, peer-to-peer, and IPC patterns (`ipcMemoryPool`). |
| `python/Utilities/` | Shared helpers imported by some samples. |
**How to run:** The top-level CMake build does **not** compile these samples. For each sample, use a Python 3.10+ environment with a matching [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads) (the samples target CUDA 13.x and document exact package pins in `requirements.txt`):
```bash
cd python/<category>/<sampleName>
pip install -r requirements.txt
python <sampleScript>.py
```
Use each sample’s `README.md` for prerequisites, CLI options, and expected output.
## Install Samples
### Installation Path Structure
The installation system automatically organizes samples into a structured directory layout based on:
- **Target Architecture**: ${CMAKE_SYSTEM_PROCESSOR}, e.g. `x64`, `aarch64`, `amd64`, etc.
- **Target OS**: `linux`, `windows`, `darwin`, `qnx`
- **Build Type**: `release`, `debug`, etc.
The default installation path is: `build/bin/${TARGET_ARCH}/${TARGET_OS}/${BUILD_TYPE}`
**Examples:**
- Linux x86_64 Release: `build/bin/x64/linux/release`
- Linux aarch64 Release: `build/bin/aarch64/linux/release`
- Windows amd64 Release: `build/bin/amd64/windows/release`
### Customizing Installation Paths
You can customize the installation location using CMake variables during the configuration step:
- `CMAKE_INSTALL_PREFIX`: Changes the root installation directory (default: `build/bin`)
```
cmake -DCMAKE_INSTALL_PREFIX=/custom/path ..
```
This will install to: `/custom/path/${TARGET_ARCH}/${TARGET_OS}/${BUILD_TYPE}`
- `CUDA_SAMPLES_INSTALL_DIR`: Specifies the exact final installation directory (overrides the structured path)
```
cmake -DCUDA_SAMPLES_INSTALL_DIR=/exact/install/path ..
```
### Install Samples on Linux
**Prerequisites:** You must first configure the project with CMake as described in the [Building CUDA Samples - Linux](#linux) or [Building]section.
After configuring and building, install the samples:
```
cd build/
make install
```
### Install Samples on Windows
**Prerequisites:** You must first configure the project with CMake as described in the [Building CUDA Samples - Windows](#windows) section.
#### Using Command Line
After configuring with CMake, build and install from the `x64 Native Tools Command Prompt for VS`:
```cmd
cd build
cmake --build . --config Release
cmake --install . --config Release
```
**Note:** Replace `Release` with `Debug` if you want to install debug builds. For multi-configuration generators (like Visual Studio), the `--config` flag determines which build type to install.
#### Using Visual Studio IDE
Alternatively, open the generated solution file `CUDA_Samples.sln` in Visual Studio:
1. Select the desired configuration (`Release` or `Debug`)
2. Build the solution (F7 or Build > Build Solution)
3. Right-click on the `INSTALL` target under `CMakePredefinedTargets` in Solution Explorer
4. Select "Build"
## Running All Samples as Tests
It's important to note that the CUDA samples are _not_ intended as a validation suite for CUDA. They do not cover corner cases, they do not completely cover the
runtime and driver APIs, are not intended for performance benchmarking, etc. That said, it can sometimes be useful to run all of the samples as a quick sanity check and
we provide a script to do so, `run_tests.py`.
This Python3 script finds all executables in a subdirectory you choose, matching application names with command line arguments specified in `test_args.json`. It accepts
the following command line arguments:
| Switch | Purpose | Example |
| ---------- | -------------------------------------------------------------------------------------------------------------- | ----------------------- |
| --dir | Specify the root directory to search for executables (recursively) | --dir ./build/cpp |
| --config | JSON configuration file for executable arguments | --config test_args.json |
| --output | Output directory for test results (stdout saved to .txt files - directory will be created if it doesn't exist) | --output ./test |
| --args | Global arguments to pass to all executables (not currently used) | --args arg_1 arg_2 ... |
| --parallel | Number of applications to execute in parallel. | --parallel 8 |
Application configurations are loaded from `test_args.json` and matched against executable names (discarding the `.exe` extension on Windows).
The script returns 0 on success, or the first non-zero error code encountered during testing on failure. It will also print a condensed list of samples that failed, if any.
There are three primary modes of configuration:
**Skip**
An executable configured with "skip" will not be executed. These generally rely on having attached graphical displays and are not suited to this kind of automation.
Configuration example:
```json
"fluidsGL": {
"skip": true
}
```
You will see:
```
Skipping fluidsGL (marked as skip in config)
```
**Single Run**
For executables to run one time only with arguments, specify each argument as a list entry. Each entry in the JSON file will be appended to the command line, separated
by a space.
All applications execute from their current directory, so all paths are relative to the application's location.
Note that if an application needs no arguments, this entry is optional. An executable found without a matching entry in the JSON will just run as `./application` from its
current directory.
Configuration example:
```json
"ptxgen": {
"args": [
"test.ll",
"-arch=compute_75"
]
}
```
You will see:
```
Running ptxgen
Command: ./ptxgen test.ll -arch=compute_75
Test completed with return code 0
```
**Multiple Runs**
For executables to run multiple times with different command line arguments, specify any number of sets of args within a "runs" list.
As with single runs, all applications execute from their current directory, so all paths are relative to the application's location.
Configuration example:
```json
"recursiveGaussian": {
"runs": [
{
"args": [
"-sigma=10",
"-file=data/ref_10.ppm"
]
},
{
"args": [
"-sigma=14",
"-file=data/ref_14.ppm"
]
},
{
"args": [
"-sigma=18",
"-file=data/ref_18.ppm"
]
},
{
"args": [
"-sigma=22",
"-file=data/ref_22.ppm"
]
}
]
}
```
You will see:
```
Running recursiveGaussian (run 1/4)
Command: ./recursiveGaussian -sigma=10 -file=data/ref_10.ppm
Test completed with return code 0
Running recursiveGaussian (run 2/4)
Command: ./recursiveGaussian -sigma=14 -file=data/ref_14.ppm
Test completed with return code 0
Running recursiveGaussian (run 3/4)
Command: ./recursiveGaussian -sigma=18 -file=data/ref_18.ppm
Test completed with return code 0
Running recursiveGaussian (run 4/4)
Command: ./recursiveGaussian -sigma=22 -file=data/ref_22.ppm
Test completed with return code 0
```
### Example Usage
Here is an example set of commands to build and test all of the samples.
First, build:
```bash
mkdir build
cd build
cmake ..
make -j$(nproc)
```
Now, return to the samples root directory and run the test script:
```bash
cd ..
python3 run_tests.py --output ./test --dir ./build/cpp --config test_args.json
```
If all applications run successfully, you will see something similar to this (the specific number of samples will depend on your build type
and system configuration):
```
Test Summary:
Ran 199 test runs for 180 executables.
All test runs passed!
```
If some samples fail, you will see something like this:
```
Test Summary:
Ran 199 test runs for 180 executables.
Failed runs (2):
bicubicTexture (run 1/5): Failed (code 1)
Mandelbrot (run 1/2): Failed (code 1)
```
You can inspect the stdout logs in the output directory (generally `APM_<application_name>.txt` or `APM_<application_name>.run<n>.txt`) to help
determine what may have gone wrong from the output logs. Please file issues against the samples repository if you believe a sample is failing
incorrectly on your system.
## Samples list ## Samples list
### [0. Introduction](./cpp/0_Introduction/README.md) ### Samples by OS
Basic CUDA samples for beginners that illustrate key concepts with using CUDA and CUDA runtime APIs.
### [1. Utilities](./cpp/1_Utilities/README.md) #### Linux
Utility samples that demonstrate how to query device capabilities and measure GPU/CPU bandwidth. **[simpleAssert_nvrtc](./Samples/simpleAssert_nvrtc)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** | **[graphMemoryFootprint](./Samples/graphMemoryFootprint)** | **[MC_EstimatePiQ](./Samples/MC_EstimatePiQ)** |
---|---|---|---|
**[reductionMultiBlockCG](./Samples/reductionMultiBlockCG)** | **[cuDLAStandaloneMode](./Samples/cuDLAStandaloneMode)** | **[conjugateGradientPrecond](./Samples/conjugateGradientPrecond)** | **[ptxjit](./Samples/ptxjit)** |
**[threadMigration](./Samples/threadMigration)** | **[EGLStream_CUDA_CrossGPU](./Samples/EGLStream_CUDA_CrossGPU)** | **[threadFenceReduction](./Samples/threadFenceReduction)** | **[simpleAtomicIntrinsics_nvrtc](./Samples/simpleAtomicIntrinsics_nvrtc)** |
**[shfl_scan](./Samples/shfl_scan)** | **[clock](./Samples/clock)** | **[binaryPartitionCG](./Samples/binaryPartitionCG)** | **[MC_EstimatePiP](./Samples/MC_EstimatePiP)** |
**[transpose](./Samples/transpose)** | **[simpleMultiCopy](./Samples/simpleMultiCopy)** | **[cuDLAErrorReporting](./Samples/cuDLAErrorReporting)** | **[concurrentKernels](./Samples/concurrentKernels)** |
**[streamOrderedAllocationIPC](./Samples/streamOrderedAllocationIPC)** | **[UnifiedMemoryPerf](./Samples/UnifiedMemoryPerf)** | **[cppIntegration](./Samples/cppIntegration)** | **[bf16TensorCoreGemm](./Samples/bf16TensorCoreGemm)** |
**[cuSolverSp_LinearSolver](./Samples/cuSolverSp_LinearSolver)** | **[nvJPEG](./Samples/nvJPEG)** | **[watershedSegmentationNPP](./Samples/watershedSegmentationNPP)** | **[MC_EstimatePiInlineP](./Samples/MC_EstimatePiInlineP)** |
**[vulkanImageCUDA](./Samples/vulkanImageCUDA)** | **[boxFilterNPP](./Samples/boxFilterNPP)** | **[simpleAssert](./Samples/simpleAssert)** | **[simpleTemplates](./Samples/simpleTemplates)** |
**[cuHook](./Samples/cuHook)** | **[simpleCUDA2GL](./Samples/simpleCUDA2GL)** | **[matrixMul](./Samples/matrixMul)** | **[quasirandomGenerator_nvrtc](./Samples/quasirandomGenerator_nvrtc)** |
**[nvJPEG_encoder](./Samples/nvJPEG_encoder)** | **[simpleCudaGraphs](./Samples/simpleCudaGraphs)** | **[simpleTextureDrv](./Samples/simpleTextureDrv)** | **[simpleCUFFT](./Samples/simpleCUFFT)** |
**[simpleCallback](./Samples/simpleCallback)** | **[batchCUBLAS](./Samples/batchCUBLAS)** | **[cudaTensorCoreGemm](./Samples/cudaTensorCoreGemm)** | **[simpleAtomicIntrinsics](./Samples/simpleAtomicIntrinsics)** |
**[newdelete](./Samples/newdelete)** | **[bicubicTexture](./Samples/bicubicTexture)** | **[dxtc](./Samples/dxtc)** | **[cudaOpenMP](./Samples/cudaOpenMP)** |
**[cdpBezierTessellation](./Samples/cdpBezierTessellation)** | **[randomFog](./Samples/randomFog)** | **[bilateralFilter](./Samples/bilateralFilter)** | **[conjugateGradient](./Samples/conjugateGradient)** |
**[particles](./Samples/particles)** | **[NV12toBGRandResize](./Samples/NV12toBGRandResize)** | **[Mandelbrot](./Samples/Mandelbrot)** | **[binomialOptions_nvrtc](./Samples/binomialOptions_nvrtc)** |
**[cudaNvSci](./Samples/cudaNvSci)** | **[mergeSort](./Samples/mergeSort)** | **[HSOpticalFlow](./Samples/HSOpticalFlow)** | **[immaTensorCoreGemm](./Samples/immaTensorCoreGemm)** |
**[convolutionTexture](./Samples/convolutionTexture)** | **[simpleVulkan](./Samples/simpleVulkan)** | **[simpleCUBLAS](./Samples/simpleCUBLAS)** | **[jacobiCudaGraphs](./Samples/jacobiCudaGraphs)** |
**[eigenvalues](./Samples/eigenvalues)** | **[streamOrderedAllocationP2P](./Samples/streamOrderedAllocationP2P)** | **[cuSolverSp_LowlevelCholesky](./Samples/cuSolverSp_LowlevelCholesky)** | **[topologyQuery](./Samples/topologyQuery)** |
**[dmmaTensorCoreGemm](./Samples/dmmaTensorCoreGemm)** | **[volumeRender](./Samples/volumeRender)** | **[stereoDisparity](./Samples/stereoDisparity)** | **[simpleTexture](./Samples/simpleTexture)** |
**[simpleStreams](./Samples/simpleStreams)** | **[smokeParticles](./Samples/smokeParticles)** | **[simpleMultiGPU](./Samples/simpleMultiGPU)** | **[deviceQueryDrv](./Samples/deviceQueryDrv)** |
**[fastWalshTransform](./Samples/fastWalshTransform)** | **[quasirandomGenerator](./Samples/quasirandomGenerator)** | **[vectorAddMMAP](./Samples/vectorAddMMAP)** | **[MersenneTwisterGP11213](./Samples/MersenneTwisterGP11213)** |
**[conjugateGradientUM](./Samples/conjugateGradientUM)** | **[simpleVoteIntrinsics_nvrtc](./Samples/simpleVoteIntrinsics_nvrtc)** | **[simpleLayeredTexture](./Samples/simpleLayeredTexture)** | **[streamOrderedAllocation](./Samples/streamOrderedAllocation)** |
**[cuSolverDn_LinearSolver](./Samples/cuSolverDn_LinearSolver)** | **[cudaCompressibleMemory](./Samples/cudaCompressibleMemory)** | **[matrixMulCUBLAS](./Samples/matrixMulCUBLAS)** | **[histEqualizationNPP](./Samples/histEqualizationNPP)** |
**[simpleAWBarrier](./Samples/simpleAWBarrier)** | **[recursiveGaussian](./Samples/recursiveGaussian)** | **[imageDenoising](./Samples/imageDenoising)** | **[FunctionPointers](./Samples/FunctionPointers)** |
**[simpleGL](./Samples/simpleGL)** | **[segmentationTreeThrust](./Samples/segmentationTreeThrust)** | **[scalarProd](./Samples/scalarProd)** | **[SobolQRNG](./Samples/SobolQRNG)** |
**[p2pBandwidthLatencyTest](./Samples/p2pBandwidthLatencyTest)** | **[conjugateGradientCudaGraphs](./Samples/conjugateGradientCudaGraphs)** | **[simplePitchLinearTexture](./Samples/simplePitchLinearTexture)** | **[freeImageInteropNPP](./Samples/freeImageInteropNPP)** |
**[template](./Samples/template)** | **[dwtHaar1D](./Samples/dwtHaar1D)** | **[postProcessGL](./Samples/postProcessGL)** | **[BlackScholes](./Samples/BlackScholes)** |
**[volumeFiltering](./Samples/volumeFiltering)** | **[simpleCUFFT_callback](./Samples/simpleCUFFT_callback)** | **[UnifiedMemoryStreams](./Samples/UnifiedMemoryStreams)** | **[globalToShmemAsyncCopy](./Samples/globalToShmemAsyncCopy)** |
**[deviceQuery](./Samples/deviceQuery)** | **[simpleHyperQ](./Samples/simpleHyperQ)** | **[systemWideAtomics](./Samples/systemWideAtomics)** | **[cuSolverSp_LowlevelQR](./Samples/cuSolverSp_LowlevelQR)** |
**[inlinePTX](./Samples/inlinePTX)** | **[matrixMulDrv](./Samples/matrixMulDrv)** | **[EGLStream_CUDA_Interop](./Samples/EGLStream_CUDA_Interop)** | **[cuDLAHybridMode](./Samples/cuDLAHybridMode)** |
**[asyncAPI](./Samples/asyncAPI)** | **[MC_EstimatePiInlineQ](./Samples/MC_EstimatePiInlineQ)** | **[scan](./Samples/scan)** | **[simpleCooperativeGroups](./Samples/simpleCooperativeGroups)** |
**[conjugateGradientMultiBlockCG](./Samples/conjugateGradientMultiBlockCG)** | **[simpleTemplates_nvrtc](./Samples/simpleTemplates_nvrtc)** | **[simpleTexture3D](./Samples/simpleTexture3D)** | **[lineOfSight](./Samples/lineOfSight)** |
**[simpleZeroCopy](./Samples/simpleZeroCopy)** | **[binomialOptions](./Samples/binomialOptions)** | **[simpleAttributes](./Samples/simpleAttributes)** | **[bindlessTexture](./Samples/bindlessTexture)** |
**[simpleCUFFT_2d_MGPU](./Samples/simpleCUFFT_2d_MGPU)** | **[simplePrintf](./Samples/simplePrintf)** | **[batchedLabelMarkersAndLabelCompressionNPP](./Samples/batchedLabelMarkersAndLabelCompressionNPP)** | **[histogram](./Samples/histogram)** |
**[matrixMulDynlinkJIT](./Samples/matrixMulDynlinkJIT)** | **[simpleP2P](./Samples/simpleP2P)** | **[conjugateGradientMultiDeviceCG](./Samples/conjugateGradientMultiDeviceCG)** | **[vectorAddDrv](./Samples/vectorAddDrv)** |
**[sortingNetworks](./Samples/sortingNetworks)** | **[alignedTypes](./Samples/alignedTypes)** | **[inlinePTX_nvrtc](./Samples/inlinePTX_nvrtc)** | **[simpleCubemapTexture](./Samples/simpleCubemapTexture)** |
**[simpleIPC](./Samples/simpleIPC)** | **[simpleVulkanMMAP](./Samples/simpleVulkanMMAP)** | **[radixSortThrust](./Samples/radixSortThrust)** | **[MonteCarloMultiGPU](./Samples/MonteCarloMultiGPU)** |
**[cudaNvSciNvMedia](./Samples/cudaNvSciNvMedia)** | **[vectorAdd](./Samples/vectorAdd)** | **[cdpSimplePrint](./Samples/cdpSimplePrint)** | **[FilterBorderControlNPP](./Samples/FilterBorderControlNPP)** |
**[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** | **[convolutionSeparable](./Samples/convolutionSeparable)** | **[nbody](./Samples/nbody)** | **[simpleCUBLAS_LU](./Samples/simpleCUBLAS_LU)** |
**[simpleSeparateCompilation](./Samples/simpleSeparateCompilation)** | **[c++11_cuda](./Samples/c++11_cuda)** | **[fluidsGL](./Samples/fluidsGL)** | **[bandwidthTest](./Samples/bandwidthTest)** |
**[clock_nvrtc](./Samples/clock_nvrtc)** | **[graphMemoryNodes](./Samples/graphMemoryNodes)** | **[cdpQuadtree](./Samples/cdpQuadtree)** | **[interval](./Samples/interval)** |
**[boxFilter](./Samples/boxFilter)** | **[matrixMul_nvrtc](./Samples/matrixMul_nvrtc)** | **[simpleCUBLASXT](./Samples/simpleCUBLASXT)** | **[cppOverload](./Samples/cppOverload)** |
**[marchingCubes](./Samples/marchingCubes)** | **[cuSolverRf](./Samples/cuSolverRf)** | **[BlackScholes_nvrtc](./Samples/BlackScholes_nvrtc)** | **[cdpAdvancedQuicksort](./Samples/cdpAdvancedQuicksort)** |
**[tf32TensorCoreGemm](./Samples/tf32TensorCoreGemm)** | **[cdpSimpleQuicksort](./Samples/cdpSimpleQuicksort)** | **[simpleOccupancy](./Samples/simpleOccupancy)** | **[simpleSurfaceWrite](./Samples/simpleSurfaceWrite)** |
**[simpleCUFFT_MGPU](./Samples/simpleCUFFT_MGPU)** | **[simpleDrvRuntime](./Samples/simpleDrvRuntime)** | **[convolutionFFT2D](./Samples/convolutionFFT2D)** | **[reduction](./Samples/reduction)** |
**[memMapIPCDrv](./Samples/memMapIPCDrv)** | **[SobelFilter](./Samples/SobelFilter)** | **[dct8x8](./Samples/dct8x8)** | **[fp16ScalarProduct](./Samples/fp16ScalarProduct)** |
**[FDTD3d](./Samples/FDTD3d)** | **[oceanFFT](./Samples/oceanFFT)** | **[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[StreamPriorities](./Samples/StreamPriorities)** |
**[cannyEdgeDetectorNPP](./Samples/cannyEdgeDetectorNPP)** | **[MC_SingleAsianOptionP](./Samples/MC_SingleAsianOptionP)** | **[simpleMPI](./Samples/simpleMPI)** |
### [2. Concepts and Techniques](./cpp/2_Concepts_and_Techniques/README.md) #### Windows
Samples that demonstrate CUDA related concepts and common problem solving techniques. **[simpleAssert_nvrtc](./Samples/simpleAssert_nvrtc)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** | **[graphMemoryFootprint](./Samples/graphMemoryFootprint)** | **[MC_EstimatePiQ](./Samples/MC_EstimatePiQ)** |
---|---|---|---|
### [3. CUDA Features](./cpp/3_CUDA_Features/README.md) **[reductionMultiBlockCG](./Samples/reductionMultiBlockCG)** | **[conjugateGradientPrecond](./Samples/conjugateGradientPrecond)** | **[ptxjit](./Samples/ptxjit)** | **[threadMigration](./Samples/threadMigration)** |
Samples that demonstrate CUDA Features (Cooperative Groups, CUDA Dynamic Parallelism, CUDA Graphs etc). **[threadFenceReduction](./Samples/threadFenceReduction)** | **[simpleAtomicIntrinsics_nvrtc](./Samples/simpleAtomicIntrinsics_nvrtc)** | **[shfl_scan](./Samples/shfl_scan)** | **[clock](./Samples/clock)** |
**[binaryPartitionCG](./Samples/binaryPartitionCG)** | **[MC_EstimatePiP](./Samples/MC_EstimatePiP)** | **[transpose](./Samples/transpose)** | **[simpleMultiCopy](./Samples/simpleMultiCopy)** |
### [4. CUDA Libraries](./cpp/4_CUDA_Libraries/README.md) **[concurrentKernels](./Samples/concurrentKernels)** | **[UnifiedMemoryPerf](./Samples/UnifiedMemoryPerf)** | **[cppIntegration](./Samples/cppIntegration)** | **[bf16TensorCoreGemm](./Samples/bf16TensorCoreGemm)** |
Samples that demonstrate how to use CUDA platform libraries (NPP, NVJPEG, NVGRAPH cuBLAS, cuFFT, cuSPARSE, cuSOLVER and cuRAND). **[cuSolverSp_LinearSolver](./Samples/cuSolverSp_LinearSolver)** | **[nvJPEG](./Samples/nvJPEG)** | **[watershedSegmentationNPP](./Samples/watershedSegmentationNPP)** | **[MC_EstimatePiInlineP](./Samples/MC_EstimatePiInlineP)** |
**[simpleD3D10](./Samples/simpleD3D10)** | **[vulkanImageCUDA](./Samples/vulkanImageCUDA)** | **[fluidsD3D9](./Samples/fluidsD3D9)** | **[boxFilterNPP](./Samples/boxFilterNPP)** |
### [5. Domain Specific](./cpp/5_Domain_Specific/README.md) **[simpleAssert](./Samples/simpleAssert)** | **[simpleTemplates](./Samples/simpleTemplates)** | **[simpleCUDA2GL](./Samples/simpleCUDA2GL)** | **[matrixMul](./Samples/matrixMul)** |
Samples that are specific to domain (Graphics, Finance, Image Processing). **[quasirandomGenerator_nvrtc](./Samples/quasirandomGenerator_nvrtc)** | **[nvJPEG_encoder](./Samples/nvJPEG_encoder)** | **[simpleCudaGraphs](./Samples/simpleCudaGraphs)** | **[simpleTextureDrv](./Samples/simpleTextureDrv)** |
**[simpleCUFFT](./Samples/simpleCUFFT)** | **[simpleCallback](./Samples/simpleCallback)** | **[SLID3D10Texture](./Samples/SLID3D10Texture)** | **[batchCUBLAS](./Samples/batchCUBLAS)** |
### [6. Performance](./cpp/6_Performance/README.md) **[cudaTensorCoreGemm](./Samples/cudaTensorCoreGemm)** | **[simpleAtomicIntrinsics](./Samples/simpleAtomicIntrinsics)** | **[newdelete](./Samples/newdelete)** | **[bicubicTexture](./Samples/bicubicTexture)** |
Samples that demonstrate performance optimization. **[dxtc](./Samples/dxtc)** | **[cudaOpenMP](./Samples/cudaOpenMP)** | **[cdpBezierTessellation](./Samples/cdpBezierTessellation)** | **[randomFog](./Samples/randomFog)** |
**[bilateralFilter](./Samples/bilateralFilter)** | **[conjugateGradient](./Samples/conjugateGradient)** | **[particles](./Samples/particles)** | **[NV12toBGRandResize](./Samples/NV12toBGRandResize)** |
### [7. libNVVM](./cpp/7_libNVVM/README.md) **[Mandelbrot](./Samples/Mandelbrot)** | **[binomialOptions_nvrtc](./Samples/binomialOptions_nvrtc)** | **[simpleD3D10RenderTarget](./Samples/simpleD3D10RenderTarget)** | **[mergeSort](./Samples/mergeSort)** |
Samples that demonstrate the use of libNVVVM and NVVM IR. **[HSOpticalFlow](./Samples/HSOpticalFlow)** | **[immaTensorCoreGemm](./Samples/immaTensorCoreGemm)** | **[convolutionTexture](./Samples/convolutionTexture)** | **[simpleVulkan](./Samples/simpleVulkan)** |
**[simpleD3D9Texture](./Samples/simpleD3D9Texture)** | **[simpleCUBLAS](./Samples/simpleCUBLAS)** | **[jacobiCudaGraphs](./Samples/jacobiCudaGraphs)** | **[eigenvalues](./Samples/eigenvalues)** |
### [8. Platform Specific](./cpp/8_Platform_Specific/Tegra/README.md) **[simpleD3D10Texture](./Samples/simpleD3D10Texture)** | **[streamOrderedAllocationP2P](./Samples/streamOrderedAllocationP2P)** | **[cuSolverSp_LowlevelCholesky](./Samples/cuSolverSp_LowlevelCholesky)** | **[topologyQuery](./Samples/topologyQuery)** |
Samples that are specific to certain platforms (Tegra, cuDLA, NvMedia, NvSci, OpenGL ES). **[dmmaTensorCoreGemm](./Samples/dmmaTensorCoreGemm)** | **[volumeRender](./Samples/volumeRender)** | **[stereoDisparity](./Samples/stereoDisparity)** | **[simpleTexture](./Samples/simpleTexture)** |
**[simpleStreams](./Samples/simpleStreams)** | **[smokeParticles](./Samples/smokeParticles)** | **[simpleMultiGPU](./Samples/simpleMultiGPU)** | **[deviceQueryDrv](./Samples/deviceQueryDrv)** |
**[fastWalshTransform](./Samples/fastWalshTransform)** | **[quasirandomGenerator](./Samples/quasirandomGenerator)** | **[vectorAddMMAP](./Samples/vectorAddMMAP)** | **[MersenneTwisterGP11213](./Samples/MersenneTwisterGP11213)** |
**[conjugateGradientUM](./Samples/conjugateGradientUM)** | **[simpleVoteIntrinsics_nvrtc](./Samples/simpleVoteIntrinsics_nvrtc)** | **[simpleLayeredTexture](./Samples/simpleLayeredTexture)** | **[streamOrderedAllocation](./Samples/streamOrderedAllocation)** |
**[cuSolverDn_LinearSolver](./Samples/cuSolverDn_LinearSolver)** | **[cudaCompressibleMemory](./Samples/cudaCompressibleMemory)** | **[matrixMulCUBLAS](./Samples/matrixMulCUBLAS)** | **[histEqualizationNPP](./Samples/histEqualizationNPP)** |
**[simpleAWBarrier](./Samples/simpleAWBarrier)** | **[recursiveGaussian](./Samples/recursiveGaussian)** | **[imageDenoising](./Samples/imageDenoising)** | **[FunctionPointers](./Samples/FunctionPointers)** |
**[simpleGL](./Samples/simpleGL)** | **[segmentationTreeThrust](./Samples/segmentationTreeThrust)** | **[scalarProd](./Samples/scalarProd)** | **[SobolQRNG](./Samples/SobolQRNG)** |
**[p2pBandwidthLatencyTest](./Samples/p2pBandwidthLatencyTest)** | **[conjugateGradientCudaGraphs](./Samples/conjugateGradientCudaGraphs)** | **[simplePitchLinearTexture](./Samples/simplePitchLinearTexture)** | **[freeImageInteropNPP](./Samples/freeImageInteropNPP)** |
**[template](./Samples/template)** | **[dwtHaar1D](./Samples/dwtHaar1D)** | **[simpleD3D11Texture](./Samples/simpleD3D11Texture)** | **[postProcessGL](./Samples/postProcessGL)** |
**[BlackScholes](./Samples/BlackScholes)** | **[simpleD3D11](./Samples/simpleD3D11)** | **[volumeFiltering](./Samples/volumeFiltering)** | **[UnifiedMemoryStreams](./Samples/UnifiedMemoryStreams)** |
**[globalToShmemAsyncCopy](./Samples/globalToShmemAsyncCopy)** | **[deviceQuery](./Samples/deviceQuery)** | **[simpleHyperQ](./Samples/simpleHyperQ)** | **[cuSolverSp_LowlevelQR](./Samples/cuSolverSp_LowlevelQR)** |
**[inlinePTX](./Samples/inlinePTX)** | **[matrixMulDrv](./Samples/matrixMulDrv)** | **[asyncAPI](./Samples/asyncAPI)** | **[MC_EstimatePiInlineQ](./Samples/MC_EstimatePiInlineQ)** |
**[scan](./Samples/scan)** | **[simpleCooperativeGroups](./Samples/simpleCooperativeGroups)** | **[conjugateGradientMultiBlockCG](./Samples/conjugateGradientMultiBlockCG)** | **[simpleTemplates_nvrtc](./Samples/simpleTemplates_nvrtc)** |
**[simpleTexture3D](./Samples/simpleTexture3D)** | **[lineOfSight](./Samples/lineOfSight)** | **[simpleZeroCopy](./Samples/simpleZeroCopy)** | **[binomialOptions](./Samples/binomialOptions)** |
**[simpleAttributes](./Samples/simpleAttributes)** | **[bindlessTexture](./Samples/bindlessTexture)** | **[simpleD3D12](./Samples/simpleD3D12)** | **[simpleCUFFT_2d_MGPU](./Samples/simpleCUFFT_2d_MGPU)** |
**[simplePrintf](./Samples/simplePrintf)** | **[batchedLabelMarkersAndLabelCompressionNPP](./Samples/batchedLabelMarkersAndLabelCompressionNPP)** | **[histogram](./Samples/histogram)** | **[matrixMulDynlinkJIT](./Samples/matrixMulDynlinkJIT)** |
**[simpleP2P](./Samples/simpleP2P)** | **[conjugateGradientMultiDeviceCG](./Samples/conjugateGradientMultiDeviceCG)** | **[vectorAddDrv](./Samples/vectorAddDrv)** | **[sortingNetworks](./Samples/sortingNetworks)** |
**[alignedTypes](./Samples/alignedTypes)** | **[inlinePTX_nvrtc](./Samples/inlinePTX_nvrtc)** | **[simpleCubemapTexture](./Samples/simpleCubemapTexture)** | **[simpleIPC](./Samples/simpleIPC)** |
**[simpleVulkanMMAP](./Samples/simpleVulkanMMAP)** | **[radixSortThrust](./Samples/radixSortThrust)** | **[MonteCarloMultiGPU](./Samples/MonteCarloMultiGPU)** | **[vectorAdd](./Samples/vectorAdd)** |
**[VFlockingD3D10](./Samples/VFlockingD3D10)** | **[simpleD3D9](./Samples/simpleD3D9)** | **[cdpSimplePrint](./Samples/cdpSimplePrint)** | **[FilterBorderControlNPP](./Samples/FilterBorderControlNPP)** |
**[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** | **[convolutionSeparable](./Samples/convolutionSeparable)** | **[nbody](./Samples/nbody)** | **[simpleCUBLAS_LU](./Samples/simpleCUBLAS_LU)** |
**[simpleSeparateCompilation](./Samples/simpleSeparateCompilation)** | **[c++11_cuda](./Samples/c++11_cuda)** | **[fluidsGL](./Samples/fluidsGL)** | **[bandwidthTest](./Samples/bandwidthTest)** |
**[clock_nvrtc](./Samples/clock_nvrtc)** | **[graphMemoryNodes](./Samples/graphMemoryNodes)** | **[cdpQuadtree](./Samples/cdpQuadtree)** | **[interval](./Samples/interval)** |
**[boxFilter](./Samples/boxFilter)** | **[matrixMul_nvrtc](./Samples/matrixMul_nvrtc)** | **[simpleCUBLASXT](./Samples/simpleCUBLASXT)** | **[cppOverload](./Samples/cppOverload)** |
**[marchingCubes](./Samples/marchingCubes)** | **[cuSolverRf](./Samples/cuSolverRf)** | **[BlackScholes_nvrtc](./Samples/BlackScholes_nvrtc)** | **[cdpAdvancedQuicksort](./Samples/cdpAdvancedQuicksort)** |
**[tf32TensorCoreGemm](./Samples/tf32TensorCoreGemm)** | **[cdpSimpleQuicksort](./Samples/cdpSimpleQuicksort)** | **[simpleOccupancy](./Samples/simpleOccupancy)** | **[simpleSurfaceWrite](./Samples/simpleSurfaceWrite)** |
**[simpleCUFFT_MGPU](./Samples/simpleCUFFT_MGPU)** | **[simpleDrvRuntime](./Samples/simpleDrvRuntime)** | **[convolutionFFT2D](./Samples/convolutionFFT2D)** | **[reduction](./Samples/reduction)** |
**[memMapIPCDrv](./Samples/memMapIPCDrv)** | **[SobelFilter](./Samples/SobelFilter)** | **[dct8x8](./Samples/dct8x8)** | **[fp16ScalarProduct](./Samples/fp16ScalarProduct)** |
**[FDTD3d](./Samples/FDTD3d)** | **[oceanFFT](./Samples/oceanFFT)** | **[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[cannyEdgeDetectorNPP](./Samples/cannyEdgeDetectorNPP)** |
**[MC_SingleAsianOptionP](./Samples/MC_SingleAsianOptionP)** | **[simpleMPI](./Samples/simpleMPI)** |
## Dependencies ## Dependencies
@@ -492,7 +296,7 @@ These third-party dependencies are required by some CUDA samples. If available,
FreeImage is an open source imaging library. FreeImage can usually be installed on Linux using your distribution's package manager system. FreeImage can also be downloaded from the FreeImage website. FreeImage is an open source imaging library. FreeImage can usually be installed on Linux using your distribution's package manager system. FreeImage can also be downloaded from the FreeImage website.
To set up FreeImage on a Windows system, extract the FreeImage DLL distribution into the folder `./Common/FreeImage/Dist/x64` such that it contains the .h and .lib files. Copy the .dll file to the Release/ Debug/ execution folder or pass the FreeImage folder when cmake configuring with the `-DFreeImage_INCLUDE_DIR` and `-DFreeImage_LIBRARY` options. To set up FreeImage on a Windows system, extract the FreeImage DLL distribution into the folder `../../Common/FreeImage/Dist/x64` such that it contains the .h, .dll, and .lib files.
#### Message Passing Interface #### Message Passing Interface
@@ -518,27 +322,13 @@ OpenGL is a graphics library used for 2D and 3D rendering. On systems which supp
OpenGL ES is an embedded systems graphics library used for 2D and 3D rendering. On systems which support OpenGL ES, NVIDIA's OpenGL ES implementation is provided with the CUDA Driver. OpenGL ES is an embedded systems graphics library used for 2D and 3D rendering. On systems which support OpenGL ES, NVIDIA's OpenGL ES implementation is provided with the CUDA Driver.
#### Freeglut
Freeglut is an open-source software library that serves as a replacement for the original OpenGL Utility Toolkit (GLUT). Its primary purpose is to make it easier for developers to create and manage windows containing OpenGL contexts, as well as handle input from devices like the mouse, keyboard, and joystick, across a wide range of platforms. To set up Freeglut on a Windowson on ARM system, you need to download the source from [Freeglut website](https://freeglut.sourceforge.net/), build freeglut on your system, and copy the freeglut.lib into the folder `./Common/lib/x64` and copy the freeglut.dll file into the `./bin/win64/${BUILD_TYPE}` execution folder.
#### Vulkan #### Vulkan
Vulkan is a low-overhead, cross-platform 3D graphics and compute API. Vulkan targets high-performance realtime 3D graphics applications such as video games and interactive media across all platforms. On systems which support Vulkan, NVIDIA's Vulkan implementation is provided with the CUDA Driver. For building and running Vulkan applications one needs to install the [Vulkan SDK](https://www.lunarg.com/vulkan-sdk/). Vulkan is a low-overhead, cross-platform 3D graphics and compute API. Vulkan targets high-performance realtime 3D graphics applications such as video games and interactive media across all platforms. On systems which support Vulkan, NVIDIA's Vulkan implementation is provided with the CUDA Driver. For building and running Vulkan applications one needs to install the [Vulkan SDK](https://www.lunarg.com/vulkan-sdk/).
#### GLEW
GLEW (OpenGL Extension Wrangler Library) is a cross-platform, open-source C/C++ library designed to simplify the process of using modern OpenGL features and extensions. Its main function is to dynamically load OpenGL function pointers at runtime, allowing developers to access both core OpenGL functions and additional features provided by hardware vendors, known as extensions. To set up GLEW on a Windows on ARM system, you need to download the source from [GLEW website](https://glew.sourceforge.net/), build GLEW on your system, and copy the glew32.lib into the folder `./Common/lib/x64` and the glew32.dll into the `./bin/win64/${BUILD_TYPE}` execution folder.
#### GLFW
GLFW is a lightweight, open-source library designed for managing OpenGL, OpenGL ES, and Vulkan contexts. It simplifies the process of creating and managing windows, handling user input (keyboard, mouse, and joystick), and working with multiple monitors in a cross-platform manner.
To set up GLFW on a Windows system, Download the pre-built binaries from [GLFW website](https://www.glfw.org/download.html) and extract the zip file into the folder, pass the GLFW include header folder as `-DGLFW_INCLUDE_DIR` and lib folder as `-DGLFW_LIB_DIR` for cmake configuring.
#### OpenMP #### OpenMP
OpenMP is an API for multiprocessing programming. OpenMP can be installed using your Linux distribution's package manager system. It usually comes preinstalled with GCC. It can also be found at the [OpenMP website](http://openmp.org/). For compilers such as clang, make sure OpenMP is enabled when building LLVM by including `openmp` in `LLVM_ENABLE_PROJECTS`. If you use clang (from an installed prefix or directly from an LLVM build tree) with OpenMP enabled, set CMAKE_CXX_COMPILER and CMAKE_CUDA_HOST_COMPILER to that clang++ and let CMake detect OpenMP; extra OpenMP_* CMake variables are usually not needed. When using clang++ directly from an LLVM build tree, you may need to copy the generated `omp.h` into the `include/` directory under the path reported by `clang++ --print-resource-dir` before building the samples. When using an installed clang with OpenMP, if you see libomp.so: cannot open shared object file at runtime, add the directory that contains libomp.so to LD_LIBRARY_PATH (or configure it via ld.so.conf.d) so the dynamic linker can locate the OpenMP runtime. OpenMP is an API for multiprocessing programming. OpenMP can be installed using your Linux distribution's package manager system. It usually comes preinstalled with GCC. It can also be found at the [OpenMP website](http://openmp.org/).
#### Screen #### Screen
@@ -628,11 +418,6 @@ NVJPEG library provides high-performance, GPU accelerated JPEG decoding function
NVRTC (CUDA RunTime Compilation) is a runtime compilation library for CUDA C++. NVRTC (CUDA RunTime Compilation) is a runtime compilation library for CUDA C++.
#### NVJITLINK
NVJITLINK is a CUDA runtime library that links multiple GPU device code objects at runtime and enables just‑in‑time link‑time optimization (JIT LTO).
It can take device code from tools such as NVCC and NVRTC, link and optimize it on the fly, and produce the final GPU binary.For more details, see: https://docs.nvidia.com/cuda/nvjitlink/index.html
#### Stream Priorities #### Stream Priorities
Stream Priorities allows the creation of streams with specified priorities. Stream Priorities is only available on GPUs with SM architecture of 3.5 or above. Stream Priorities allows the creation of streams with specified priorities. Stream Priorities is only available on GPUs with SM architecture of 3.5 or above.
@@ -649,10 +434,6 @@ FP16 is a 16-bit floating-point format. One bit is used for the sign, five bits
NVCC support of [C++11 features](https://en.wikipedia.org/wiki/C++11). NVCC support of [C++11 features](https://en.wikipedia.org/wiki/C++11).
#### CMake
The libNVVM samples are built using [CMake](https://cmake.org/) 3.10 or later.
## Contributors Guide ## Contributors Guide
We welcome your input on issues and suggestions for samples. At this time we are not accepting contributions from the public, check back here as we evolve our contribution model. We welcome your input on issues and suggestions for samples. At this time we are not accepting contributions from the public, check back here as we evolve our contribution model.
@@ -665,9 +446,6 @@ Answers to frequently asked questions about CUDA can be found at http://develope
## References ## References
* [CUDA Programming Guide](https://docs.nvidia.com/cuda/cuda-programming-guide/index.html) * [CUDA Programming Guide](http://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html)
* [Accelerated Computing Blog](https://developer.nvidia.com/blog/?tags=accelerated-computing) * [Accelerated Computing Blog](https://developer.nvidia.com/blog/?tags=accelerated-computing)
## Attributions
* Teapot image is obtained from [Wikimedia](https://en.wikipedia.org/wiki/File:Original_Utah_Teapot.jpg) and is licensed under the Creative Commons [Attribution-Share Alike 2.0](https://creativecommons.org/licenses/by-sa/2.0/deed.en) Generic license. The image is modified for samples use cases.

View File

@@ -0,0 +1,18 @@
{
"configurations": [
{
"name": "Linux",
"includePath": [
"${workspaceFolder}/**",
"${workspaceFolder}/../../Common"
],
"defines": [],
"compilerPath": "/usr/local/cuda/bin/nvcc",
"cStandard": "gnu17",
"cppStandard": "gnu++14",
"intelliSenseMode": "linux-gcc-x64",
"configurationProvider": "ms-vscode.makefile-tools"
}
],
"version": 4
}

View File

@@ -0,0 +1,10 @@
{
"configurations": [
{
"name": "CUDA C++: Launch",
"type": "cuda-gdb",
"request": "launch",
"program": "${workspaceFolder}/BlackScholes"
}
]
}

15
Samples/BlackScholes/.vscode/tasks.json vendored Normal file
View File

@@ -0,0 +1,15 @@
{
"version": "2.0.0",
"tasks": [
{
"label": "sample",
"type": "shell",
"command": "make dbg=1",
"problemMatcher": ["$nvcc"],
"group": {
"kind": "build",
"isDefault": true
}
}
]
}

View File

@@ -0,0 +1,243 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/*
* This sample evaluates fair call and put prices for a
* given set of European options by Black-Scholes formula.
* See supplied whitepaper for more explanations.
*/
#include <helper_functions.h> // helper functions for string parsing
#include <helper_cuda.h> // helper functions CUDA error checking and initialization
////////////////////////////////////////////////////////////////////////////////
// Process an array of optN options on CPU
////////////////////////////////////////////////////////////////////////////////
extern "C" void BlackScholesCPU(float *h_CallResult, float *h_PutResult,
float *h_StockPrice, float *h_OptionStrike,
float *h_OptionYears, float Riskfree,
float Volatility, int optN);
////////////////////////////////////////////////////////////////////////////////
// Process an array of OptN options on GPU
////////////////////////////////////////////////////////////////////////////////
#include "BlackScholes_kernel.cuh"
////////////////////////////////////////////////////////////////////////////////
// Helper function, returning uniformly distributed
// random float in [low, high] range
////////////////////////////////////////////////////////////////////////////////
float RandFloat(float low, float high) {
float t = (float)rand() / (float)RAND_MAX;
return (1.0f - t) * low + t * high;
}
////////////////////////////////////////////////////////////////////////////////
// Data configuration
////////////////////////////////////////////////////////////////////////////////
const int OPT_N = 4000000;
const int NUM_ITERATIONS = 512;
const int OPT_SZ = OPT_N * sizeof(float);
const float RISKFREE = 0.02f;
const float VOLATILITY = 0.30f;
#define DIV_UP(a, b) (((a) + (b)-1) / (b))
////////////////////////////////////////////////////////////////////////////////
// Main program
////////////////////////////////////////////////////////////////////////////////
int main(int argc, char **argv) {
// Start logs
printf("[%s] - Starting...\n", argv[0]);
//'h_' prefix - CPU (host) memory space
float
// Results calculated by CPU for reference
*h_CallResultCPU,
*h_PutResultCPU,
// CPU copy of GPU results
*h_CallResultGPU, *h_PutResultGPU,
// CPU instance of input data
*h_StockPrice, *h_OptionStrike, *h_OptionYears;
//'d_' prefix - GPU (device) memory space
float
// Results calculated by GPU
*d_CallResult,
*d_PutResult,
// GPU instance of input data
*d_StockPrice, *d_OptionStrike, *d_OptionYears;
double delta, ref, sum_delta, sum_ref, max_delta, L1norm, gpuTime;
StopWatchInterface *hTimer = NULL;
int i;
findCudaDevice(argc, (const char **)argv);
sdkCreateTimer(&hTimer);
printf("Initializing data...\n");
printf("...allocating CPU memory for options.\n");
h_CallResultCPU = (float *)malloc(OPT_SZ);
h_PutResultCPU = (float *)malloc(OPT_SZ);
h_CallResultGPU = (float *)malloc(OPT_SZ);
h_PutResultGPU = (float *)malloc(OPT_SZ);
h_StockPrice = (float *)malloc(OPT_SZ);
h_OptionStrike = (float *)malloc(OPT_SZ);
h_OptionYears = (float *)malloc(OPT_SZ);
printf("...allocating GPU memory for options.\n");
checkCudaErrors(cudaMalloc((void **)&d_CallResult, OPT_SZ));
checkCudaErrors(cudaMalloc((void **)&d_PutResult, OPT_SZ));
checkCudaErrors(cudaMalloc((void **)&d_StockPrice, OPT_SZ));
checkCudaErrors(cudaMalloc((void **)&d_OptionStrike, OPT_SZ));
checkCudaErrors(cudaMalloc((void **)&d_OptionYears, OPT_SZ));
printf("...generating input data in CPU mem.\n");
srand(5347);
// Generate options set
for (i = 0; i < OPT_N; i++) {
h_CallResultCPU[i] = 0.0f;
h_PutResultCPU[i] = -1.0f;
h_StockPrice[i] = RandFloat(5.0f, 30.0f);
h_OptionStrike[i] = RandFloat(1.0f, 100.0f);
h_OptionYears[i] = RandFloat(0.25f, 10.0f);
}
printf("...copying input data to GPU mem.\n");
// Copy options data to GPU memory for further processing
checkCudaErrors(
cudaMemcpy(d_StockPrice, h_StockPrice, OPT_SZ, cudaMemcpyHostToDevice));
checkCudaErrors(cudaMemcpy(d_OptionStrike, h_OptionStrike, OPT_SZ,
cudaMemcpyHostToDevice));
checkCudaErrors(
cudaMemcpy(d_OptionYears, h_OptionYears, OPT_SZ, cudaMemcpyHostToDevice));
printf("Data init done.\n\n");
printf("Executing Black-Scholes GPU kernel (%i iterations)...\n",
NUM_ITERATIONS);
checkCudaErrors(cudaDeviceSynchronize());
sdkResetTimer(&hTimer);
sdkStartTimer(&hTimer);
for (i = 0; i < NUM_ITERATIONS; i++) {
BlackScholesGPU<<<DIV_UP((OPT_N / 2), 128), 128 /*480, 128*/>>>(
(float2 *)d_CallResult, (float2 *)d_PutResult, (float2 *)d_StockPrice,
(float2 *)d_OptionStrike, (float2 *)d_OptionYears, RISKFREE, VOLATILITY,
OPT_N);
getLastCudaError("BlackScholesGPU() execution failed\n");
}
checkCudaErrors(cudaDeviceSynchronize());
sdkStopTimer(&hTimer);
gpuTime = sdkGetTimerValue(&hTimer) / NUM_ITERATIONS;
// Both call and put is calculated
printf("Options count : %i \n", 2 * OPT_N);
printf("BlackScholesGPU() time : %f msec\n", gpuTime);
printf("Effective memory bandwidth: %f GB/s\n",
((double)(5 * OPT_N * sizeof(float)) * 1E-9) / (gpuTime * 1E-3));
printf("Gigaoptions per second : %f \n\n",
((double)(2 * OPT_N) * 1E-9) / (gpuTime * 1E-3));
printf(
"BlackScholes, Throughput = %.4f GOptions/s, Time = %.5f s, Size = %u "
"options, NumDevsUsed = %u, Workgroup = %u\n",
(((double)(2.0 * OPT_N) * 1.0E-9) / (gpuTime * 1.0E-3)), gpuTime * 1e-3,
(2 * OPT_N), 1, 128);
printf("\nReading back GPU results...\n");
// Read back GPU results to compare them to CPU results
checkCudaErrors(cudaMemcpy(h_CallResultGPU, d_CallResult, OPT_SZ,
cudaMemcpyDeviceToHost));
checkCudaErrors(
cudaMemcpy(h_PutResultGPU, d_PutResult, OPT_SZ, cudaMemcpyDeviceToHost));
printf("Checking the results...\n");
printf("...running CPU calculations.\n\n");
// Calculate options values on CPU
BlackScholesCPU(h_CallResultCPU, h_PutResultCPU, h_StockPrice, h_OptionStrike,
h_OptionYears, RISKFREE, VOLATILITY, OPT_N);
printf("Comparing the results...\n");
// Calculate max absolute difference and L1 distance
// between CPU and GPU results
sum_delta = 0;
sum_ref = 0;
max_delta = 0;
for (i = 0; i < OPT_N; i++) {
ref = h_CallResultCPU[i];
delta = fabs(h_CallResultCPU[i] - h_CallResultGPU[i]);
if (delta > max_delta) {
max_delta = delta;
}
sum_delta += delta;
sum_ref += fabs(ref);
}
L1norm = sum_delta / sum_ref;
printf("L1 norm: %E\n", L1norm);
printf("Max absolute error: %E\n\n", max_delta);
printf("Shutting down...\n");
printf("...releasing GPU memory.\n");
checkCudaErrors(cudaFree(d_OptionYears));
checkCudaErrors(cudaFree(d_OptionStrike));
checkCudaErrors(cudaFree(d_StockPrice));
checkCudaErrors(cudaFree(d_PutResult));
checkCudaErrors(cudaFree(d_CallResult));
printf("...releasing CPU memory.\n");
free(h_OptionYears);
free(h_OptionStrike);
free(h_StockPrice);
free(h_PutResultGPU);
free(h_CallResultGPU);
free(h_PutResultCPU);
free(h_CallResultCPU);
sdkDeleteTimer(&hTimer);
printf("Shutdown done.\n");
printf("\n[BlackScholes] - Test Summary\n");
if (L1norm > 1e-6) {
printf("Test failed!\n");
exit(EXIT_FAILURE);
}
printf(
"\nNOTE: The CUDA Samples are not meant for performance measurements. "
"Results may vary when GPU Boost is enabled.\n\n");
printf("Test passed\n");
exit(EXIT_SUCCESS);
}

View File

@@ -0,0 +1,86 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <math.h>
////////////////////////////////////////////////////////////////////////////////
// Polynomial approximation of cumulative normal distribution function
////////////////////////////////////////////////////////////////////////////////
static double CND(double d) {
const double A1 = 0.31938153;
const double A2 = -0.356563782;
const double A3 = 1.781477937;
const double A4 = -1.821255978;
const double A5 = 1.330274429;
const double RSQRT2PI = 0.39894228040143267793994605993438;
double K = 1.0 / (1.0 + 0.2316419 * fabs(d));
double cnd = RSQRT2PI * exp(-0.5 * d * d) *
(K * (A1 + K * (A2 + K * (A3 + K * (A4 + K * A5)))));
if (d > 0) cnd = 1.0 - cnd;
return cnd;
}
////////////////////////////////////////////////////////////////////////////////
// Black-Scholes formula for both call and put
////////////////////////////////////////////////////////////////////////////////
static void BlackScholesBodyCPU(float &callResult, float &putResult,
float Sf, // Stock price
float Xf, // Option strike
float Tf, // Option years
float Rf, // Riskless rate
float Vf // Volatility rate
) {
double S = Sf, X = Xf, T = Tf, R = Rf, V = Vf;
double sqrtT = sqrt(T);
double d1 = (log(S / X) + (R + 0.5 * V * V) * T) / (V * sqrtT);
double d2 = d1 - V * sqrtT;
double CNDD1 = CND(d1);
double CNDD2 = CND(d2);
// Calculate Call and Put simultaneously
double expRT = exp(-R * T);
callResult = (float)(S * CNDD1 - X * expRT * CNDD2);
putResult = (float)(X * expRT * (1.0 - CNDD2) - S * (1.0 - CNDD1));
}
////////////////////////////////////////////////////////////////////////////////
// Process an array of optN options
////////////////////////////////////////////////////////////////////////////////
extern "C" void BlackScholesCPU(float *h_CallResult, float *h_PutResult,
float *h_StockPrice, float *h_OptionStrike,
float *h_OptionYears, float Riskfree,
float Volatility, int optN) {
for (int opt = 0; opt < optN; opt++)
BlackScholesBodyCPU(h_CallResult[opt], h_PutResult[opt], h_StockPrice[opt],
h_OptionStrike[opt], h_OptionYears[opt], Riskfree,
Volatility);
}

View File

@@ -0,0 +1,106 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
////////////////////////////////////////////////////////////////////////////////
// Polynomial approximation of cumulative normal distribution function
////////////////////////////////////////////////////////////////////////////////
__device__ inline float cndGPU(float d) {
const float A1 = 0.31938153f;
const float A2 = -0.356563782f;
const float A3 = 1.781477937f;
const float A4 = -1.821255978f;
const float A5 = 1.330274429f;
const float RSQRT2PI = 0.39894228040143267793994605993438f;
float K = __fdividef(1.0f, (1.0f + 0.2316419f * fabsf(d)));
float cnd = RSQRT2PI * __expf(-0.5f * d * d) *
(K * (A1 + K * (A2 + K * (A3 + K * (A4 + K * A5)))));
if (d > 0) cnd = 1.0f - cnd;
return cnd;
}
////////////////////////////////////////////////////////////////////////////////
// Black-Scholes formula for both call and put
////////////////////////////////////////////////////////////////////////////////
__device__ inline void BlackScholesBodyGPU(float &CallResult, float &PutResult,
float S, // Stock price
float X, // Option strike
float T, // Option years
float R, // Riskless rate
float V // Volatility rate
) {
float sqrtT, expRT;
float d1, d2, CNDD1, CNDD2;
sqrtT = __fdividef(1.0F, rsqrtf(T));
d1 = __fdividef(__logf(S / X) + (R + 0.5f * V * V) * T, V * sqrtT);
d2 = d1 - V * sqrtT;
CNDD1 = cndGPU(d1);
CNDD2 = cndGPU(d2);
// Calculate Call and Put simultaneously
expRT = __expf(-R * T);
CallResult = S * CNDD1 - X * expRT * CNDD2;
PutResult = X * expRT * (1.0f - CNDD2) - S * (1.0f - CNDD1);
}
////////////////////////////////////////////////////////////////////////////////
// Process an array of optN options on GPU
////////////////////////////////////////////////////////////////////////////////
__launch_bounds__(128) __global__
void BlackScholesGPU(float2 *__restrict d_CallResult,
float2 *__restrict d_PutResult,
float2 *__restrict d_StockPrice,
float2 *__restrict d_OptionStrike,
float2 *__restrict d_OptionYears, float Riskfree,
float Volatility, int optN) {
////Thread index
// const int tid = blockDim.x * blockIdx.x + threadIdx.x;
////Total number of threads in execution grid
// const int THREAD_N = blockDim.x * gridDim.x;
const int opt = blockDim.x * blockIdx.x + threadIdx.x;
// Calculating 2 options per thread to increase ILP (instruction level
// parallelism)
if (opt < (optN / 2)) {
float callResult1, callResult2;
float putResult1, putResult2;
BlackScholesBodyGPU(callResult1, putResult1, d_StockPrice[opt].x,
d_OptionStrike[opt].x, d_OptionYears[opt].x, Riskfree,
Volatility);
BlackScholesBodyGPU(callResult2, putResult2, d_StockPrice[opt].y,
d_OptionStrike[opt].y, d_OptionYears[opt].y, Riskfree,
Volatility);
d_CallResult[opt] = make_float2(callResult1, callResult2);
d_PutResult[opt] = make_float2(putResult1, putResult2);
}
}

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2017
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "BlackScholes", "BlackScholes_vs2017.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,114 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>BlackScholes_vs2017</RootNamespace>
<ProjectName>BlackScholes</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<PropertyGroup Condition="'$(WindowsTargetPlatformVersion)'==''">
<LatestTargetPlatformVersion>$([Microsoft.Build.Utilities.ToolLocationHelper]::GetLatestSDKTargetPlatformVersion('Windows', '10.0'))</LatestTargetPlatformVersion>
<WindowsTargetPlatformVersion Condition="'$(WindowsTargetPlatformVersion)' == ''">$(LatestTargetPlatformVersion)</WindowsTargetPlatformVersion>
<TargetPlatformVersion>$(WindowsTargetPlatformVersion)</TargetPlatformVersion>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v141</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/BlackScholes.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,sm_35;compute_37,sm_37;compute_50,sm_50;compute_52,sm_52;compute_60,sm_60;compute_61,sm_61;compute_70,sm_70;compute_75,sm_75;compute_80,sm_80;compute_86,sm_86;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<CudaCompile Include="BlackScholes.cu" />
<ClCompile Include="BlackScholes_gold.cpp" />
<None Include="BlackScholes_kernel.cuh" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2019
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "BlackScholes", "BlackScholes_vs2019.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,110 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>BlackScholes_vs2019</RootNamespace>
<ProjectName>BlackScholes</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v142</PlatformToolset>
<WindowsTargetPlatformVersion>10.0</WindowsTargetPlatformVersion>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/BlackScholes.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,sm_35;compute_37,sm_37;compute_50,sm_50;compute_52,sm_52;compute_60,sm_60;compute_61,sm_61;compute_70,sm_70;compute_75,sm_75;compute_80,sm_80;compute_86,sm_86;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<CudaCompile Include="BlackScholes.cu" />
<ClCompile Include="BlackScholes_gold.cpp" />
<None Include="BlackScholes_kernel.cuh" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,365 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
NVCCFLAGS += -g -G
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
SAMPLE_ENABLED := 1
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
#Detect if installed version of GCC supports required C++11
ifeq ($(TARGET_OS),linux)
empty :=
space := $(empty) $(empty)
GCCVERSIONSTRING := $(shell expr `$(HOST_COMPILER) -dumpversion`)
#Create version number without "."
GCCVERSION := $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f1 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f2 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f3 -d.)
# Make sure the version number has at least 3 decimals
GCCVERSION += 00
# Remove spaces from the version number
GCCVERSION := $(subst $(space),$(empty),$(GCCVERSION))
#$(warning $(GCCVERSION))
IS_MIN_VERSION := $(shell expr `echo $(GCCVERSION)` \>= 48000)
ifeq ($(IS_MIN_VERSION), 1)
$(info >>> GCC Version is greater or equal to 4.8.0 <<<)
else
$(info >>> Waiving build. Minimum GCC version required is 4.8.0<<<)
SAMPLE_ENABLED := 0
endif
endif
# Gencode arguments
ifeq ($(TARGET_ARCH),$(filter $(TARGET_ARCH),armv7l aarch64))
SMS ?= 35 37 50 52 60 61 70 72 75 80 86
else
SMS ?= 35 37 50 52 60 61 70 75 80 86
endif
ifeq ($(SMS),)
$(info >>> WARNING - no SM architectures have been specified - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ifeq ($(GENCODE_FLAGS),)
# Generate SASS code for each SM architecture listed in $(SMS)
$(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
# Generate PTX code from the highest SM architecture in $(SMS) to guarantee forward-compatibility
HIGHEST_SM := $(lastword $(sort $(SMS)))
ifneq ($(HIGHEST_SM),)
GENCODE_FLAGS += -gencode arch=compute_$(HIGHEST_SM),code=compute_$(HIGHEST_SM)
endif
endif
ALL_CCFLAGS += -maxrregcount=16 --threads 0 --std=c++11
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: BlackScholes
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
BlackScholes.o:BlackScholes.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
BlackScholes_gold.o:BlackScholes_gold.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
BlackScholes: BlackScholes.o BlackScholes_gold.o
$(EXEC) $(NVCC) $(ALL_LDFLAGS) $(GENCODE_FLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./BlackScholes
clean:
rm -f BlackScholes BlackScholes.o BlackScholes_gold.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/BlackScholes
clobber: clean

View File

@@ -0,0 +1,71 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>BlackScholes</name>
<cflags>
<flag>-maxrregcount=16</flag>
</cflags>
<description><![CDATA[This sample evaluates fair call and put prices for a given set of European options by Black-Scholes formula.]]></description>
<devicecompilation>whole</devicecompilation>
<includepaths>
<path>./</path>
<path>../</path>
<path>../../Common</path>
</includepaths>
<keyconcepts>
<concept level="basic">Computational Finance</concept>
</keyconcepts>
<keywords>
<keyword>CUDA</keyword>
<keyword>Computational Finance</keyword>
<keyword>option pricing</keyword>
<keyword>Black-Scholes</keyword>
</keywords>
<libraries>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>BlackScholes.cu</primary_file>
<scopes>
<scope>1:CUDA Basic Topics</scope>
<scope>3:Computational Finance</scope>
</scopes>
<sm-arch>sm35</sm-arch>
<sm-arch>sm37</sm-arch>
<sm-arch>sm50</sm-arch>
<sm-arch>sm52</sm-arch>
<sm-arch>sm60</sm-arch>
<sm-arch>sm61</sm-arch>
<sm-arch>sm70</sm-arch>
<sm-arch>sm72</sm-arch>
<sm-arch>sm75</sm-arch>
<sm-arch>sm80</sm-arch>
<sm-arch>sm86</sm-arch>
<supported_envs>
<env>
<arch>x86_64</arch>
<platform>linux</platform>
</env>
<env>
<platform>windows7</platform>
</env>
<env>
<arch>x86_64</arch>
<platform>macosx</platform>
</env>
<env>
<arch>arm</arch>
</env>
<env>
<arch>ppc64le</arch>
<platform>linux</platform>
</env>
</supported_envs>
<supported_sm_architectures>
<include>all</include>
</supported_sm_architectures>
<title>Black-Scholes Option Pricing</title>
<type>exe</type>
<whitepaper>doc\BlackScholes.pdf</whitepaper>
</entry>

View File

@@ -0,0 +1,67 @@
# BlackScholes - Black-Scholes Option Pricing
## Description
This sample evaluates fair call and put prices for a given set of European options by Black-Scholes formula.
## Key Concepts
Computational Finance
## Supported SM Architectures
[SM 3.5 ](https://developer.nvidia.com/cuda-gpus) [SM 3.7 ](https://developer.nvidia.com/cuda-gpus) [SM 5.0 ](https://developer.nvidia.com/cuda-gpus) [SM 5.2 ](https://developer.nvidia.com/cuda-gpus) [SM 6.0 ](https://developer.nvidia.com/cuda-gpus) [SM 6.1 ](https://developer.nvidia.com/cuda-gpus) [SM 7.0 ](https://developer.nvidia.com/cuda-gpus) [SM 7.2 ](https://developer.nvidia.com/cuda-gpus) [SM 7.5 ](https://developer.nvidia.com/cuda-gpus) [SM 8.0 ](https://developer.nvidia.com/cuda-gpus) [SM 8.6 ](https://developer.nvidia.com/cuda-gpus)
## Supported OSes
Linux, Windows
## Supported CPU Architecture
x86_64, ppc64le, armv7l
## CUDA APIs involved
## Prerequisites
Download and install the [CUDA Toolkit 11.5](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
## Build and Run
### Windows
The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format:
```
*_vs<version>.sln - for Visual Studio <version>
```
Each individual sample has its own set of solution files in its directory:
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
> **Note:** Some samples require that the Microsoft DirectX SDK (June 2010 or newer) be installed and that the VC++ directory paths are properly set up (**Tools > Options...**). Check DirectX Dependencies section for details."
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, ppc64le, armv7l.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=ppc64le` <br/> `$ make TARGET_ARCH=armv7l` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,18 @@
{
"configurations": [
{
"name": "Linux",
"includePath": [
"${workspaceFolder}/**",
"${workspaceFolder}/../../Common"
],
"defines": [],
"compilerPath": "/usr/local/cuda/bin/nvcc",
"cStandard": "gnu17",
"cppStandard": "gnu++14",
"intelliSenseMode": "linux-gcc-x64",
"configurationProvider": "ms-vscode.makefile-tools"
}
],
"version": 4
}

View File

@@ -0,0 +1,10 @@
{
"configurations": [
{
"name": "CUDA C++: Launch",
"type": "cuda-gdb",
"request": "launch",
"program": "${workspaceFolder}/BlackScholes_nvrtc"
}
]
}

View File

@@ -0,0 +1,15 @@
{
"version": "2.0.0",
"tasks": [
{
"label": "sample",
"type": "shell",
"command": "make dbg=1",
"problemMatcher": ["$nvcc"],
"group": {
"kind": "build",
"isDefault": true
}
}
]
}

View File

@@ -0,0 +1,269 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/*
* This sample evaluates fair call and put prices for a
* given set of European options by Black-Scholes formula.
* See supplied whitepaper for more explanations.
*/
#include <cuda_runtime.h>
#include <nvrtc_helper.h>
#include <helper_functions.h> // helper functions for string parsing
////////////////////////////////////////////////////////////////////////////////
// Process an array of optN options on CPU
////////////////////////////////////////////////////////////////////////////////
extern "C" void BlackScholesCPU(float *h_CallResult, float *h_PutResult,
float *h_StockPrice, float *h_OptionStrike,
float *h_OptionYears, float Riskfree,
float Volatility, int optN);
////////////////////////////////////////////////////////////////////////////////
// Process an array of OptN options on GPU
////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////
// Helper function, returning uniformly distributed
// random float in [low, high] range
////////////////////////////////////////////////////////////////////////////////
float RandFloat(float low, float high) {
float t = (float)rand() / (float)RAND_MAX;
return (1.0f - t) * low + t * high;
}
////////////////////////////////////////////////////////////////////////////////
// Data configuration
////////////////////////////////////////////////////////////////////////////////
const int OPT_N = 4000000;
const int NUM_ITERATIONS = 512;
const int OPT_SZ = OPT_N * sizeof(float);
const float RISKFREE = 0.02f;
const float VOLATILITY = 0.30f;
#define DIV_UP(a, b) (((a) + (b)-1) / (b))
////////////////////////////////////////////////////////////////////////////////
// Main program
////////////////////////////////////////////////////////////////////////////////
int main(int argc, char **argv) {
// Start logs
printf("[%s] - Starting...\n", argv[0]);
//'h_' prefix - CPU (host) memory space
float
// Results calculated by CPU for reference
*h_CallResultCPU,
*h_PutResultCPU,
// CPU copy of GPU results
*h_CallResultGPU, *h_PutResultGPU,
// CPU instance of input data
*h_StockPrice, *h_OptionStrike, *h_OptionYears;
//'d_' prefix - GPU (device) memory space
CUdeviceptr
// Results calculated by GPU
d_CallResult,
d_PutResult,
// GPU instance of input data
d_StockPrice, d_OptionStrike, d_OptionYears;
double delta, ref, sum_delta, sum_ref, max_delta, L1norm, gpuTime;
StopWatchInterface *hTimer = NULL;
int i;
sdkCreateTimer(&hTimer);
printf("Initializing data...\n");
printf("...allocating CPU memory for options.\n");
h_CallResultCPU = (float *)malloc(OPT_SZ);
h_PutResultCPU = (float *)malloc(OPT_SZ);
h_CallResultGPU = (float *)malloc(OPT_SZ);
h_PutResultGPU = (float *)malloc(OPT_SZ);
h_StockPrice = (float *)malloc(OPT_SZ);
h_OptionStrike = (float *)malloc(OPT_SZ);
h_OptionYears = (float *)malloc(OPT_SZ);
char *cubin, *kernel_file;
size_t cubinSize;
kernel_file = sdkFindFilePath("BlackScholes_kernel.cuh", argv[0]);
// Compile the kernel BlackScholes_kernel.
compileFileToCUBIN(kernel_file, argc, argv, &cubin, &cubinSize, 0);
CUmodule module = loadCUBIN(cubin, argc, argv);
CUfunction kernel_addr;
checkCudaErrors(cuModuleGetFunction(&kernel_addr, module, "BlackScholesGPU"));
printf("...allocating GPU memory for options.\n");
checkCudaErrors(cuMemAlloc(&d_CallResult, OPT_SZ));
checkCudaErrors(cuMemAlloc(&d_PutResult, OPT_SZ));
checkCudaErrors(cuMemAlloc(&d_StockPrice, OPT_SZ));
checkCudaErrors(cuMemAlloc(&d_OptionStrike, OPT_SZ));
checkCudaErrors(cuMemAlloc(&d_OptionYears, OPT_SZ));
printf("...generating input data in CPU mem.\n");
srand(5347);
// Generate options set
for (i = 0; i < OPT_N; i++) {
h_CallResultCPU[i] = 0.0f;
h_PutResultCPU[i] = -1.0f;
h_StockPrice[i] = RandFloat(5.0f, 30.0f);
h_OptionStrike[i] = RandFloat(1.0f, 100.0f);
h_OptionYears[i] = RandFloat(0.25f, 10.0f);
}
printf("...copying input data to GPU mem.\n");
// Copy options data to GPU memory for further processing
checkCudaErrors(cuMemcpyHtoD(d_StockPrice, h_StockPrice, OPT_SZ));
checkCudaErrors(cuMemcpyHtoD(d_OptionStrike, h_OptionStrike, OPT_SZ));
checkCudaErrors(cuMemcpyHtoD(d_OptionYears, h_OptionYears, OPT_SZ));
printf("Data init done.\n\n");
printf("Executing Black-Scholes GPU kernel (%i iterations)...\n",
NUM_ITERATIONS);
sdkResetTimer(&hTimer);
sdkStartTimer(&hTimer);
dim3 cudaBlockSize(128, 1, 1);
dim3 cudaGridSize(DIV_UP(OPT_N / 2, 128), 1, 1);
float risk = RISKFREE;
float volatility = VOLATILITY;
int optval = OPT_N;
void *arr[] = {(void *)&d_CallResult, (void *)&d_PutResult,
(void *)&d_StockPrice, (void *)&d_OptionStrike,
(void *)&d_OptionYears, (void *)&risk,
(void *)&volatility, (void *)&optval};
for (i = 0; i < NUM_ITERATIONS; i++) {
checkCudaErrors(cuLaunchKernel(kernel_addr, cudaGridSize.x, cudaGridSize.y,
cudaGridSize.z, /* grid dim */
cudaBlockSize.x, cudaBlockSize.y,
cudaBlockSize.z, /* block dim */
0, 0, /* shared mem, stream */
&arr[0], /* arguments */
0));
}
checkCudaErrors(cuCtxSynchronize());
sdkStopTimer(&hTimer);
gpuTime = sdkGetTimerValue(&hTimer) / NUM_ITERATIONS;
// Both call and put is calculated
printf("Options count : %i \n", 2 * OPT_N);
printf("BlackScholesGPU() time : %f msec\n", gpuTime);
printf("Effective memory bandwidth: %f GB/s\n",
((double)(5 * OPT_N * sizeof(float)) * 1E-9) / (gpuTime * 1E-3));
printf("Gigaoptions per second : %f \n\n",
((double)(2 * OPT_N) * 1E-9) / (gpuTime * 1E-3));
printf(
"BlackScholes, Throughput = %.4f GOptions/s, Time = %.5f s, Size = %u "
"options, NumDevsUsed = %u, Workgroup = %u\n",
(((double)(2.0 * OPT_N) * 1.0E-9) / (gpuTime * 1.0E-3)), gpuTime * 1e-3,
(2 * OPT_N), 1, 128);
printf("\nReading back GPU results...\n");
// Read back GPU results to compare them to CPU results
checkCudaErrors(cuMemcpyDtoH(h_CallResultGPU, d_CallResult, OPT_SZ));
checkCudaErrors(cuMemcpyDtoH(h_PutResultGPU, d_PutResult, OPT_SZ));
printf("Checking the results...\n");
printf("...running CPU calculations.\n\n");
// Calculate options values on CPU
BlackScholesCPU(h_CallResultCPU, h_PutResultCPU, h_StockPrice, h_OptionStrike,
h_OptionYears, RISKFREE, VOLATILITY, OPT_N);
printf("Comparing the results...\n");
// Calculate max absolute difference and L1 distance
// between CPU and GPU results
sum_delta = 0;
sum_ref = 0;
max_delta = 0;
for (i = 0; i < OPT_N; i++) {
ref = h_CallResultCPU[i];
delta = fabs(h_CallResultCPU[i] - h_CallResultGPU[i]);
if (delta > max_delta) {
max_delta = delta;
}
sum_delta += delta;
sum_ref += fabs(ref);
}
L1norm = sum_delta / sum_ref;
printf("L1 norm: %E\n", L1norm);
printf("Max absolute error: %E\n\n", max_delta);
printf("Shutting down...\n");
printf("...releasing GPU memory.\n");
checkCudaErrors(cuMemFree(d_OptionYears));
checkCudaErrors(cuMemFree(d_OptionStrike));
checkCudaErrors(cuMemFree(d_StockPrice));
checkCudaErrors(cuMemFree(d_PutResult));
checkCudaErrors(cuMemFree(d_CallResult));
printf("...releasing CPU memory.\n");
free(h_OptionYears);
free(h_OptionStrike);
free(h_StockPrice);
free(h_PutResultGPU);
free(h_CallResultGPU);
free(h_PutResultCPU);
free(h_CallResultCPU);
sdkDeleteTimer(&hTimer);
printf("Shutdown done.\n");
printf("\n[%s] - Test Summary\n", argv[0]);
if (L1norm > 1e-6) {
printf("Test failed!\n");
exit(EXIT_FAILURE);
}
printf("Test passed\n");
exit(EXIT_SUCCESS);
}

View File

@@ -0,0 +1,88 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <math.h>
///////////////////////////////////////////////////////////////////////////////
// Polynomial approximation of cumulative normal distribution function
///////////////////////////////////////////////////////////////////////////////
static double CND(double d) {
const double A1 = 0.31938153;
const double A2 = -0.356563782;
const double A3 = 1.781477937;
const double A4 = -1.821255978;
const double A5 = 1.330274429;
const double RSQRT2PI = 0.39894228040143267793994605993438;
double K = 1.0 / (1.0 + 0.2316419 * fabs(d));
double cnd = RSQRT2PI * exp(-0.5 * d * d) *
(K * (A1 + K * (A2 + K * (A3 + K * (A4 + K * A5)))));
if (d > 0) cnd = 1.0 - cnd;
return cnd;
}
///////////////////////////////////////////////////////////////////////////////
// Black-Scholes formula for both call and put
///////////////////////////////////////////////////////////////////////////////
static void BlackScholesBodyCPU(float &callResult, float &putResult,
float Sf, // Stock price
float Xf, // Option strike
float Tf, // Option years
float Rf, // Riskless rate
float Vf // Volatility rate
) {
double S = Sf, X = Xf, T = Tf, R = Rf, V = Vf;
double sqrtT = sqrt(T);
double d1 = (log(S / X) + (R + 0.5 * V * V) * T) / (V * sqrtT);
double d2 = d1 - V * sqrtT;
double CNDD1 = CND(d1);
double CNDD2 = CND(d2);
// Calculate Call and Put simultaneously
double expRT = exp(-R * T);
callResult = (float)(S * CNDD1 - X * expRT * CNDD2);
putResult = (float)(X * expRT * (1.0 - CNDD2) - S * (1.0 - CNDD1));
}
////////////////////////////////////////////////////////////////////////////////
// Process an array of optN options
////////////////////////////////////////////////////////////////////////////////
extern "C" void BlackScholesCPU(float *h_CallResult, float *h_PutResult,
float *h_StockPrice, float *h_OptionStrike,
float *h_OptionYears, float Riskfree,
float Volatility, int optN) {
for (int opt = 0; opt < optN; opt++)
BlackScholesBodyCPU(h_CallResult[opt], h_PutResult[opt], h_StockPrice[opt],
h_OptionStrike[opt], h_OptionYears[opt], Riskfree,
Volatility);
}

View File

@@ -0,0 +1,103 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
///////////////////////////////////////////////////////////////////////////////
// Polynomial approximation of cumulative normal distribution function
///////////////////////////////////////////////////////////////////////////////
__device__ inline float cndGPU(float d) {
const float A1 = 0.31938153f;
const float A2 = -0.356563782f;
const float A3 = 1.781477937f;
const float A4 = -1.821255978f;
const float A5 = 1.330274429f;
const float RSQRT2PI = 0.39894228040143267793994605993438f;
float K = __fdividef(1.0f, (1.0f + 0.2316419f * fabsf(d)));
float cnd = RSQRT2PI * __expf(-0.5f * d * d) *
(K * (A1 + K * (A2 + K * (A3 + K * (A4 + K * A5)))));
if (d > 0) cnd = 1.0f - cnd;
return cnd;
}
///////////////////////////////////////////////////////////////////////////////
// Black-Scholes formula for both call and put
///////////////////////////////////////////////////////////////////////////////
__device__ inline void BlackScholesBodyGPU(float &CallResult, float &PutResult,
float S, // Stock price
float X, // Option strike
float T, // Option years
float R, // Riskless rate
float V // Volatility rate
) {
float sqrtT, expRT;
float d1, d2, CNDD1, CNDD2;
sqrtT = __fdividef(1.0F, rsqrtf(T));
d1 = __fdividef(__logf(S / X) + (R + 0.5f * V * V) * T, V * sqrtT);
d2 = d1 - V * sqrtT;
CNDD1 = cndGPU(d1);
CNDD2 = cndGPU(d2);
// Calculate Call and Put simultaneously
expRT = __expf(-R * T);
CallResult = S * CNDD1 - X * expRT * CNDD2;
PutResult = X * expRT * (1.0f - CNDD2) - S * (1.0f - CNDD1);
}
////////////////////////////////////////////////////////////////////////////////
// Process an array of optN options on GPU
////////////////////////////////////////////////////////////////////////////////
extern "C" __launch_bounds__(128) __global__
void BlackScholesGPU(float2 *__restrict d_CallResult,
float2 *__restrict d_PutResult,
float2 *__restrict d_StockPrice,
float2 *__restrict d_OptionStrike,
float2 *__restrict d_OptionYears, float Riskfree,
float Volatility, int optN) {
////Thread index
const int opt = blockDim.x * blockIdx.x + threadIdx.x;
// Calculating 2 options per thread to increase ILP (instruction level
// parallelism)
if (opt < (optN / 2)) {
float callResult1, callResult2;
float putResult1, putResult2;
BlackScholesBodyGPU(callResult1, putResult1, d_StockPrice[opt].x,
d_OptionStrike[opt].x, d_OptionYears[opt].x, Riskfree,
Volatility);
BlackScholesBodyGPU(callResult2, putResult2, d_StockPrice[opt].y,
d_OptionStrike[opt].y, d_OptionYears[opt].y, Riskfree,
Volatility);
d_CallResult[opt] = make_float2(callResult1, callResult2);
d_PutResult[opt] = make_float2(putResult1, putResult2);
}
}

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2017
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "BlackScholes_nvrtc", "BlackScholes_nvrtc_vs2017.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,114 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>BlackScholes_nvrtc_vs2017</RootNamespace>
<ProjectName>BlackScholes_nvrtc</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<PropertyGroup Condition="'$(WindowsTargetPlatformVersion)'==''">
<LatestTargetPlatformVersion>$([Microsoft.Build.Utilities.ToolLocationHelper]::GetLatestSDKTargetPlatformVersion('Windows', '10.0'))</LatestTargetPlatformVersion>
<WindowsTargetPlatformVersion Condition="'$(WindowsTargetPlatformVersion)' == ''">$(LatestTargetPlatformVersion)</WindowsTargetPlatformVersion>
<TargetPlatformVersion>$(WindowsTargetPlatformVersion)</TargetPlatformVersion>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v141</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;$(CudaToolkitIncludeDir);$(CUDA_PATH)/include;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cuda.lib;nvrtc.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/BlackScholes_nvrtc.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration></CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="BlackScholes.cpp" />
<ClCompile Include="BlackScholes_gold.cpp" />
<None Include="BlackScholes_kernel.cuh" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2019
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "BlackScholes_nvrtc", "BlackScholes_nvrtc_vs2019.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,110 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>BlackScholes_nvrtc_vs2019</RootNamespace>
<ProjectName>BlackScholes_nvrtc</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v142</PlatformToolset>
<WindowsTargetPlatformVersion>10.0</WindowsTargetPlatformVersion>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;$(CudaToolkitIncludeDir);$(CUDA_PATH)/include;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cuda.lib;nvrtc.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/BlackScholes_nvrtc.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration></CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="BlackScholes.cpp" />
<ClCompile Include="BlackScholes_gold.cpp" />
<None Include="BlackScholes_kernel.cuh" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,422 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
CCFLAGS += -g
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
UBUNTU = $(shell lsb_release -i -s 2>/dev/null | grep -i ubuntu)
SAMPLE_ENABLED := 1
# This sample is not supported on ARMv7
ifeq ($(TARGET_ARCH),armv7l)
$(info >>> WARNING - BlackScholes_nvrtc is not supported on ARMv7 - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
# This sample is not supported on QNX
ifeq ($(TARGET_OS),qnx)
$(info >>> WARNING - BlackScholes_nvrtc is not supported on QNX - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
# libNVRTC specific libraries
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -L$(CUDA_PATH)/lib -F/Library/Frameworks -framework CUDA
endif
#Detect if installed version of GCC supports required C++11
ifeq ($(TARGET_OS),linux)
empty :=
space := $(empty) $(empty)
GCCVERSIONSTRING := $(shell expr `$(HOST_COMPILER) -dumpversion`)
#Create version number without "."
GCCVERSION := $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f1 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f2 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f3 -d.)
# Make sure the version number has at least 3 decimals
GCCVERSION += 00
# Remove spaces from the version number
GCCVERSION := $(subst $(space),$(empty),$(GCCVERSION))
#$(warning $(GCCVERSION))
IS_MIN_VERSION := $(shell expr `echo $(GCCVERSION)` \>= 48000)
ifeq ($(IS_MIN_VERSION), 1)
$(info >>> GCC Version is greater or equal to 4.8.0 <<<)
else
$(info >>> Waiving build. Minimum GCC version required is 4.8.0<<<)
SAMPLE_ENABLED := 0
endif
endif
ifeq ($(TARGET_OS),darwin)
ALL_LDFLAGS += -Xcompiler -F/Library/Frameworks -Xlinker -framework -Xlinker CUDA
else
ifeq ($(TARGET_ARCH),x86_64)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/lib64/stubs
CUDA_SEARCH_PATH += $(CUDA_PATH)/targets/x86_64-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/armv7-linux-gnueabihf/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/sbsa-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/armv7-linux-androideabi/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-linux-androideabi/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/ARMv7-linux-QNX/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-qnx/lib/stubs
ifdef TARGET_OVERRIDE
CUDA_SEARCH_PATH := $(CUDA_PATH)/targets/$(TARGET_OVERRIDE)/lib/stubs
endif
endif
ifeq ($(TARGET_ARCH),ppc64le)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/ppc64le-linux/lib/stubs
endif
ifeq ($(HOST_ARCH),ppc64le)
CUDA_SEARCH_PATH += $(CUDA_PATH)/lib64/stubs
endif
CUDALIB ?= $(shell find -L $(CUDA_SEARCH_PATH) -maxdepth 1 -name libcuda.so 2> /dev/null)
ifeq ("$(CUDALIB)","")
$(info >>> WARNING - libcuda.so not found, CUDA Driver is not installed. Please re-install the driver. <<<)
SAMPLE_ENABLED := 0
else
CUDALIB := $(shell echo $(CUDALIB) | sed "s/ .*//" | sed "s/\/libcuda.so//" )
LIBRARIES += -L$(CUDALIB) -lcuda
endif
endif
ALL_CCFLAGS += --threads 0 --std=c++11
INCLUDES += -I$(CUDA_PATH)/include
LIBRARIES += -lnvrtc
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: BlackScholes_nvrtc
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
BlackScholes.o:BlackScholes.cpp
$(EXEC) $(HOST_COMPILER) $(INCLUDES) $(CCFLAGS) $(EXTRA_CCFLAGS) -o $@ -c $<
BlackScholes_gold.o:BlackScholes_gold.cpp
$(EXEC) $(HOST_COMPILER) $(INCLUDES) $(CCFLAGS) $(EXTRA_CCFLAGS) -o $@ -c $<
BlackScholes_nvrtc: BlackScholes.o BlackScholes_gold.o
$(EXEC) $(HOST_COMPILER) $(LDFLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./BlackScholes_nvrtc
clean:
rm -f BlackScholes_nvrtc BlackScholes.o BlackScholes_gold.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/BlackScholes_nvrtc
clobber: clean

View File

@@ -0,0 +1,71 @@
# BlackScholes_nvrtc - Black-Scholes Option Pricing with libNVRTC
## Description
This sample evaluates fair call and put prices for a given set of European options by Black-Scholes formula, compiling the CUDA kernels involved at runtime using NVRTC.
## Key Concepts
Computational Finance, Runtime Compilation
## Supported SM Architectures
[SM 3.5 ](https://developer.nvidia.com/cuda-gpus) [SM 3.7 ](https://developer.nvidia.com/cuda-gpus) [SM 5.0 ](https://developer.nvidia.com/cuda-gpus) [SM 5.2 ](https://developer.nvidia.com/cuda-gpus) [SM 6.0 ](https://developer.nvidia.com/cuda-gpus) [SM 6.1 ](https://developer.nvidia.com/cuda-gpus) [SM 7.0 ](https://developer.nvidia.com/cuda-gpus) [SM 7.2 ](https://developer.nvidia.com/cuda-gpus) [SM 7.5 ](https://developer.nvidia.com/cuda-gpus) [SM 8.0 ](https://developer.nvidia.com/cuda-gpus) [SM 8.6 ](https://developer.nvidia.com/cuda-gpus)
## Supported OSes
Linux, Windows
## Supported CPU Architecture
x86_64, ppc64le, aarch64
## CUDA APIs involved
## Dependencies needed to build/run
[NVRTC](../../README.md#nvrtc)
## Prerequisites
Download and install the [CUDA Toolkit 11.5](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Make sure the dependencies mentioned in [Dependencies]() section above are installed.
## Build and Run
### Windows
The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format:
```
*_vs<version>.sln - for Visual Studio <version>
```
Each individual sample has its own set of solution files in its directory:
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
> **Note:** Some samples require that the Microsoft DirectX SDK (June 2010 or newer) be installed and that the VC++ directory paths are properly set up (**Tools > Options...**). Check DirectX Dependencies section for details."
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, ppc64le, aarch64.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=ppc64le` <br/> `$ make TARGET_ARCH=aarch64` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,18 @@
{
"configurations": [
{
"name": "Linux",
"includePath": [
"${workspaceFolder}/**",
"${workspaceFolder}/../../Common"
],
"defines": [],
"compilerPath": "/usr/local/cuda/bin/nvcc",
"cStandard": "gnu17",
"cppStandard": "gnu++14",
"intelliSenseMode": "linux-gcc-x64",
"configurationProvider": "ms-vscode.makefile-tools"
}
],
"version": 4
}

View File

@@ -0,0 +1,10 @@
{
"configurations": [
{
"name": "CUDA C++: Launch",
"type": "cuda-gdb",
"request": "launch",
"program": "${workspaceFolder}/EGLStream_CUDA_CrossGPU"
}
]
}

View File

@@ -0,0 +1,15 @@
{
"version": "2.0.0",
"tasks": [
{
"label": "sample",
"type": "shell",
"command": "make dbg=1",
"problemMatcher": ["$nvcc"],
"group": {
"kind": "build",
"isDefault": true
}
}
]
}

View File

@@ -0,0 +1,453 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
NVCCFLAGS += -g -G
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
UBUNTU = $(shell lsb_release -i -s 2>/dev/null | grep -i ubuntu)
SAMPLE_ENABLED := 1
# This sample is not supported on Mac OSX
ifeq ($(TARGET_OS),darwin)
$(info >>> WARNING - EGLStream_CUDA_CrossGPU is not supported on Mac OSX - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
# This sample is not supported on android
ifeq ($(TARGET_OS),android)
$(info >>> WARNING - EGLStream_CUDA_CrossGPU is not supported on android - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
# Makefile include to help find EGL Libraries
include ./findegl.mk
# EGL specific libraries
ifneq ($(TARGET_OS),darwin)
LIBRARIES += -lEGL
endif
#Detect if installed version of GCC supports required C++11
ifeq ($(TARGET_OS),linux)
empty :=
space := $(empty) $(empty)
GCCVERSIONSTRING := $(shell expr `$(HOST_COMPILER) -dumpversion`)
#Create version number without "."
GCCVERSION := $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f1 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f2 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f3 -d.)
# Make sure the version number has at least 3 decimals
GCCVERSION += 00
# Remove spaces from the version number
GCCVERSION := $(subst $(space),$(empty),$(GCCVERSION))
#$(warning $(GCCVERSION))
IS_MIN_VERSION := $(shell expr `echo $(GCCVERSION)` \>= 48000)
ifeq ($(IS_MIN_VERSION), 1)
$(info >>> GCC Version is greater or equal to 4.8.0 <<<)
else
$(info >>> Waiving build. Minimum GCC version required is 4.8.0<<<)
SAMPLE_ENABLED := 0
endif
endif
# Gencode arguments
ifeq ($(TARGET_ARCH),$(filter $(TARGET_ARCH),armv7l aarch64))
SMS ?= 35 37 50 52 60 61 70 72 75 80 86
else
SMS ?= 35 37 50 52 60 61 70 75 80 86
endif
ifeq ($(SMS),)
$(info >>> WARNING - no SM architectures have been specified - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ifeq ($(GENCODE_FLAGS),)
# Generate SASS code for each SM architecture listed in $(SMS)
$(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
# Generate PTX code from the highest SM architecture in $(SMS) to guarantee forward-compatibility
HIGHEST_SM := $(lastword $(sort $(SMS)))
ifneq ($(HIGHEST_SM),)
GENCODE_FLAGS += -gencode arch=compute_$(HIGHEST_SM),code=compute_$(HIGHEST_SM)
endif
endif
ifeq ($(TARGET_OS),darwin)
ALL_LDFLAGS += -Xcompiler -F/Library/Frameworks -Xlinker -framework -Xlinker CUDA
else
ifeq ($(TARGET_ARCH),x86_64)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/lib64/stubs
CUDA_SEARCH_PATH += $(CUDA_PATH)/targets/x86_64-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/armv7-linux-gnueabihf/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/sbsa-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/armv7-linux-androideabi/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-linux-androideabi/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/ARMv7-linux-QNX/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-qnx/lib/stubs
ifdef TARGET_OVERRIDE
CUDA_SEARCH_PATH := $(CUDA_PATH)/targets/$(TARGET_OVERRIDE)/lib/stubs
endif
endif
ifeq ($(TARGET_ARCH),ppc64le)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/ppc64le-linux/lib/stubs
endif
ifeq ($(HOST_ARCH),ppc64le)
CUDA_SEARCH_PATH += $(CUDA_PATH)/lib64/stubs
endif
CUDALIB ?= $(shell find -L $(CUDA_SEARCH_PATH) -maxdepth 1 -name libcuda.so 2> /dev/null)
ifeq ("$(CUDALIB)","")
$(info >>> WARNING - libcuda.so not found, CUDA Driver is not installed. Please re-install the driver. <<<)
SAMPLE_ENABLED := 0
else
CUDALIB := $(shell echo $(CUDALIB) | sed "s/ .*//" | sed "s/\/libcuda.so//" )
LIBRARIES += -L$(CUDALIB) -lcuda
endif
endif
ALL_CCFLAGS += --threads 0 --std=c++11
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: EGLStream_CUDA_CrossGPU
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
cuda_consumer.o:cuda_consumer.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
cuda_producer.o:cuda_producer.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
eglstrm_common.o:eglstrm_common.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
kernel.o:kernel.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
main.o:main.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
EGLStream_CUDA_CrossGPU: cuda_consumer.o cuda_producer.o eglstrm_common.o kernel.o main.o
$(EXEC) $(NVCC) $(ALL_LDFLAGS) $(GENCODE_FLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./EGLStream_CUDA_CrossGPU
clean:
rm -f EGLStream_CUDA_CrossGPU cuda_consumer.o cuda_producer.o eglstrm_common.o kernel.o main.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/EGLStream_CUDA_CrossGPU
clobber: clean

View File

@@ -0,0 +1,78 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>EGLStream_CUDA_CrossGPU</name>
<cuda_api_list>
<driver>cuDeviceGet</driver>
<driver>cuDeviceGetAttribute</driver>
<driver>cuDeviceComputeCapability</driver>
<driver>cuDeviceGetCount</driver>
<driver>cuDeviceGetName</driver>
<driver>cuGraphicsResourceGetMappedEglFrame</driver>
<driver>cuEGLStreamConsumerAcquireFrame</driver>
<driver>cuEGLStreamConsumerReleaseFrame</driver>
<driver>cuEGLStreamProducerReturnFrame</driver>
<driver>cuEGLStreamProducerPresentFrame</driver>
<driver>cuCtxCreate</driver>
<driver>cuMemAlloc</driver>
<driver>cuMemFree</driver>
<driver>cuMemcpy3D</driver>
<driver>cuStreamCreate</driver>
<driver>cuCtxPushCurrent</driver>
<driver>cuCtxPopCurrent</driver>
<driver>cuCtxDestroy</driver>
</cuda_api_list>
<description><![CDATA[Demonstrates CUDA and EGL Streams interop, where consumer's EGL Stream is on one GPU and producer's on other and both consumer-producer are different processes.]]></description>
<devicecompilation>whole</devicecompilation>
<includepaths>
<path>./</path>
<path>../</path>
<path>../../Common</path>
</includepaths>
<keyconcepts>
<concept level="basic">EGLStreams Interop</concept>
</keyconcepts>
<keywords>
<keyword>EGL Streams</keyword>
</keywords>
<libraries>
<library os="linux">cuda</library>
<library framework="true" os="macosx">CUDA</library>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>main.cpp</primary_file>
<required_dependencies>
<dependency>EGL</dependency>
</required_dependencies>
<scopes>
<scope>1:CUDA Advanced Topics</scope>
<scope>2:Graphics Interop</scope>
</scopes>
<sm-arch>sm35</sm-arch>
<sm-arch>sm37</sm-arch>
<sm-arch>sm50</sm-arch>
<sm-arch>sm52</sm-arch>
<sm-arch>sm60</sm-arch>
<sm-arch>sm61</sm-arch>
<sm-arch>sm70</sm-arch>
<sm-arch>sm72</sm-arch>
<sm-arch>sm75</sm-arch>
<sm-arch>sm80</sm-arch>
<sm-arch>sm86</sm-arch>
<supported_envs>
<env>
<arch>x86_64</arch>
<platform>linux</platform>
</env>
<env>
<arch>arm</arch>
</env>
</supported_envs>
<supported_sm_architectures>
<include>all</include>
</supported_sm_architectures>
<title>EGLStream_CUDA_CrossGPU</title>
<type>exe</type>
</entry>

View File

@@ -0,0 +1,64 @@
# EGLStream_CUDA_CrossGPU - EGLStream_CUDA_CrossGPU
## Description
Demonstrates CUDA and EGL Streams interop, where consumer's EGL Stream is on one GPU and producer's on other and both consumer-producer are different processes.
## Key Concepts
EGLStreams Interop
## Supported SM Architectures
[SM 3.5 ](https://developer.nvidia.com/cuda-gpus) [SM 3.7 ](https://developer.nvidia.com/cuda-gpus) [SM 5.0 ](https://developer.nvidia.com/cuda-gpus) [SM 5.2 ](https://developer.nvidia.com/cuda-gpus) [SM 6.0 ](https://developer.nvidia.com/cuda-gpus) [SM 6.1 ](https://developer.nvidia.com/cuda-gpus) [SM 7.0 ](https://developer.nvidia.com/cuda-gpus) [SM 7.2 ](https://developer.nvidia.com/cuda-gpus) [SM 7.5 ](https://developer.nvidia.com/cuda-gpus) [SM 8.0 ](https://developer.nvidia.com/cuda-gpus) [SM 8.6 ](https://developer.nvidia.com/cuda-gpus)
## Supported OSes
Linux
## Supported CPU Architecture
x86_64, armv7l
## CUDA APIs involved
### [CUDA Driver API](http://docs.nvidia.com/cuda/cuda-driver-api/index.html)
cuDeviceGet, cuDeviceGetAttribute, cuDeviceComputeCapability, cuDeviceGetCount, cuDeviceGetName, cuGraphicsResourceGetMappedEglFrame, cuEGLStreamConsumerAcquireFrame, cuEGLStreamConsumerReleaseFrame, cuEGLStreamProducerReturnFrame, cuEGLStreamProducerPresentFrame, cuCtxCreate, cuMemAlloc, cuMemFree, cuMemcpy3D, cuStreamCreate, cuCtxPushCurrent, cuCtxPopCurrent, cuCtxDestroy
## Dependencies needed to build/run
[EGL](../../README.md#egl)
## Prerequisites
Download and install the [CUDA Toolkit 11.5](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Make sure the dependencies mentioned in [Dependencies]() section above are installed.
## Build and Run
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, armv7l.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=armv7l` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,258 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Simple CUDA consumer rendering sample app
//
#include <cuda_runtime.h>
#include "cuda_consumer.h"
#include "eglstrm_common.h"
#include <math.h>
#include <unistd.h>
#if defined(EXTENSION_LIST)
EXTENSION_LIST(EXTLST_EXTERN)
#endif
CUgraphicsResource cudaResource;
static int count_acq = 0;
static double acquire_time[25000] = {0}, total_time_acq = 0;
static int count_rel = 0;
static double rel_time[25000] = {0}, total_time_rel = 0;
void acquireApiStat(void);
void acquireApiStat(void) {
int i = 0;
double min = 10000000, max = 0;
double average_launch_time = 0, standard_deviation = 0;
if (count_acq == 0) return;
// lets compute the standard deviation
min = max = acquire_time[1];
average_launch_time = (total_time_acq - acquire_time[0]) / count_acq;
for (i = 1; i < count_acq; i++) {
standard_deviation += (acquire_time[i] - average_launch_time) *
(acquire_time[i] - average_launch_time);
if (acquire_time[i] < min) min = acquire_time[i];
if (acquire_time[i] > max) max = acquire_time[i];
}
standard_deviation = sqrt(standard_deviation / count_acq);
printf("acquire Avg: %lf\n", average_launch_time);
printf("acquire SD: %lf\n", standard_deviation);
printf("acquire min: %lf\n", min);
printf("acquire max: %lf\n", max);
min = max = rel_time[1];
average_launch_time = (total_time_rel - rel_time[0]) / count_rel;
for (i = 1; i < count_rel; i++) {
standard_deviation += (rel_time[i] - average_launch_time) *
(rel_time[i] - average_launch_time);
if (rel_time[i] < min) min = rel_time[i];
if (rel_time[i] > max) max = rel_time[i];
}
standard_deviation = sqrt(standard_deviation / count_rel);
printf("release Avg: %lf\n", average_launch_time);
printf("release SD: %lf\n", standard_deviation);
printf("release min: %lf\n", min);
printf("release max: %lf\n", max);
}
CUresult cudaConsumerAcquireFrame(test_cuda_consumer_s *cudaConsumer,
int frameNumber) {
CUresult cuStatus = CUDA_SUCCESS;
CUeglFrame cudaEgl;
struct timespec start, end;
EGLint streamState = 0;
double curTime;
if (!cudaConsumer) {
printf("%s: Bad parameter\n", __func__);
goto done;
}
while (1) {
if (!eglQueryStreamKHR(cudaConsumer->eglDisplay, cudaConsumer->eglStream,
EGL_STREAM_STATE_KHR, &streamState)) {
printf("Cuda Consumer: eglQueryStreamKHR EGL_STREAM_STATE_KHR failed\n");
cuStatus = CUDA_ERROR_UNKNOWN;
goto done;
}
if (streamState == EGL_STREAM_STATE_DISCONNECTED_KHR) {
printf("Cuda Consumer: EGL_STREAM_STATE_DISCONNECTED_KHR received\n");
cuStatus = CUDA_ERROR_UNKNOWN;
goto done;
}
if (streamState == EGL_STREAM_STATE_NEW_FRAME_AVAILABLE_KHR) {
break;
}
}
if (cudaConsumer->profileAPI) {
getTime(&start);
}
cuStatus =
cuEGLStreamConsumerAcquireFrame(&(cudaConsumer->cudaConn), &cudaResource,
&cudaConsumer->consCudaStream, 16000);
if (cudaConsumer->profileAPI) {
getTime(&end);
curTime = TIME_DIFF(end, start);
acquire_time[count_acq++] = curTime;
if (count_acq == 25000) count_acq = 0;
total_time_acq += curTime;
}
if (cuStatus == CUDA_SUCCESS) {
CUdeviceptr pDevPtr = 0;
cudaError_t err;
cuStatus =
cuGraphicsResourceGetMappedEglFrame(&cudaEgl, cudaResource, 0, 0);
if (cuStatus != CUDA_SUCCESS) {
printf("Cuda get resource failed with %d\n", cuStatus);
goto done;
}
pDevPtr = (CUdeviceptr)cudaEgl.frame.pPitch[0];
err = cudaConsumer_filter(cudaConsumer->consCudaStream, (char *)pDevPtr,
WIDTH * 4, HEIGHT, PROD_DATA + frameNumber,
CONS_DATA + frameNumber, frameNumber);
if (err != cudaSuccess) {
printf("Cuda Consumer: kernel failed with: %s\n",
cudaGetErrorString(err));
goto done;
}
}
done:
return cuStatus;
}
CUresult cudaConsumerReleaseFrame(test_cuda_consumer_s *cudaConsumer,
int frameNumber) {
CUresult cuStatus = CUDA_SUCCESS;
struct timespec start, end;
double curTime;
if (!cudaConsumer) {
printf("%s: Bad parameter\n", __func__);
goto done;
}
if (cudaConsumer->profileAPI) {
getTime(&start);
}
cuStatus = cuEGLStreamConsumerReleaseFrame(
&cudaConsumer->cudaConn, cudaResource, &cudaConsumer->consCudaStream);
if (cudaConsumer->profileAPI) {
getTime(&end);
curTime = TIME_DIFF(end, start);
rel_time[count_rel++] = curTime;
if (count_rel == 25000) count_rel = 0;
total_time_rel += curTime;
}
if (cuStatus != CUDA_SUCCESS) {
printf("cuEGLStreamConsumerReleaseFrame failed, status:%d\n", cuStatus);
goto done;
}
done:
return cuStatus;
}
CUresult cudaDeviceCreateConsumer(test_cuda_consumer_s *cudaConsumer) {
CUdevice device;
CUresult status = CUDA_SUCCESS;
if (CUDA_SUCCESS != (status = cuInit(0))) {
printf("Failed to initialize CUDA\n");
return status;
}
if (CUDA_SUCCESS !=
(status = cuDeviceGet(&device, cudaConsumer->cudaDevId))) {
printf("failed to get CUDA device\n");
return status;
}
if (CUDA_SUCCESS !=
(status = cuCtxCreate(&cudaConsumer->context, 0, device))) {
printf("failed to create CUDA context\n");
return status;
}
int major = 0, minor = 0;
char deviceName[256];
cuDeviceGetAttribute(&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR,
device);
cuDeviceGetAttribute(&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR,
device);
cuDeviceGetName(deviceName, 256, device);
printf(
"CUDA Consumer on GPU Device %d: \"%s\" with compute capability "
"%d.%d\n\n",
device, deviceName, major, minor);
cuCtxPopCurrent(&cudaConsumer->context);
if (major < 6) {
printf(
"EGLStream_CUDA_CrossGPU requires SM 6.0 or higher arch GPU. "
"Exiting...\n");
exit(2); // EXIT_WAIVED
}
return status;
}
CUresult cuda_consumer_init(test_cuda_consumer_s *cudaConsumer,
TestArgs *args) {
CUresult status = CUDA_SUCCESS;
int bufferSize;
cudaConsumer->charCnt = args->charCnt;
bufferSize = args->charCnt;
cudaConsumer->pCudaCopyMem = (unsigned char *)malloc(bufferSize);
if (cudaConsumer->pCudaCopyMem == NULL) {
printf("Cuda Consumer: malloc failed\n");
goto done;
}
status = cuStreamCreate(&cudaConsumer->consCudaStream, 0);
if (status != CUDA_SUCCESS) {
printf("Cuda Consumer: cuStreamCreate failed, status:%d\n", status);
goto done;
}
atexit(acquireApiStat);
done:
return status;
}
CUresult cuda_consumer_Deinit(test_cuda_consumer_s *cudaConsumer) {
if (cudaConsumer->pCudaCopyMem) {
free(cudaConsumer->pCudaCopyMem);
}
return cuEGLStreamConsumerDisconnect(&cudaConsumer->cudaConn);
}

View File

@@ -0,0 +1,66 @@
/* Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: CUDA consumer header file
//
#ifndef _CUDA_CONSUMER_H_
#define _CUDA_CONSUMER_H_
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "cudaEGL.h"
#include "eglstrm_common.h"
#include <cuda_runtime.h>
#include <cuda.h>
typedef struct _test_cuda_consumer_s {
CUcontext context;
CUeglStreamConnection cudaConn;
int cudaDevId;
EGLDisplay eglDisplay;
EGLStreamKHR eglStream;
unsigned int charCnt;
char *cudaBuf;
bool profileAPI;
unsigned char *pCudaCopyMem;
CUstream consCudaStream;
} test_cuda_consumer_s;
CUresult cuda_consumer_init(test_cuda_consumer_s *cudaConsumer, TestArgs *args);
CUresult cuda_consumer_Deinit(test_cuda_consumer_s *cudaConsumer);
CUresult cudaConsumerAcquireFrame(test_cuda_consumer_s *data, int frameNumber);
CUresult cudaConsumerReleaseFrame(test_cuda_consumer_s *data, int frameNumber);
CUresult cudaDeviceCreateConsumer(test_cuda_consumer_s *cudaConsumer);
cudaError_t cudaConsumer_filter(CUstream cStream, char *pSrc, int width,
int height, char expectedVal, char newVal,
int frameNumber);
cudaError_t cudaGetValueMismatch(void);
#endif

Some files were not shown because too many files have changed in this diff Show More