mirror of
https://github.com/NVIDIA/cuda-samples.git
synced 2026-11-18 01:48:16 +08:00
Compare commits
96 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
3f1c509650 | ||
|
|
fbc40da311 | ||
|
|
1b36fefcd5 | ||
|
|
13c2fd9717 | ||
|
|
a5267b83a5 | ||
|
|
2ab16e6d15 | ||
|
|
903c840a8c | ||
|
|
24e7c5b428 | ||
|
|
b38ed29c95 | ||
|
|
4a631c9fd6 | ||
|
|
98afcd6515 | ||
|
|
8013610840 | ||
|
|
2369d4c649 | ||
|
|
8433f89993 | ||
|
|
a9a4db666f | ||
|
|
87d03ef003 | ||
|
|
45314d7ff8 | ||
|
|
2f0a55d7dc | ||
|
|
0d61031846 | ||
|
|
225f84d433 | ||
|
|
e674cc36fe | ||
|
|
a47b422205 | ||
|
|
ce28796d6c | ||
|
|
9424c15848 | ||
|
|
9075c50a3d | ||
|
|
de5fa98e6e | ||
|
|
6c9e9d3cd2 | ||
|
|
7f5390cec3 | ||
|
|
5f6d46dfea | ||
|
|
49307463b5 | ||
|
|
9c0d5aaae6 | ||
|
|
74a9a08887 | ||
|
|
0d08748ffa | ||
|
|
27f47634a0 | ||
|
|
40d297dfe7 | ||
|
|
20b90b5063 | ||
|
|
8a9e2c830c | ||
|
|
61225e22f0 | ||
|
|
f47f06f077 | ||
|
|
0c78a3e0de | ||
|
|
8219570c15 | ||
|
|
fc88988f23 | ||
|
|
adacf1cffd | ||
|
|
3e631f467a | ||
|
|
1141dd7af4 | ||
|
|
da3b7a2b3c | ||
|
|
5987a9e9fa | ||
|
|
107f3f537f | ||
|
|
c1b03b9f81 | ||
|
|
ee0ee417c9 | ||
|
|
ebc1078379 | ||
|
|
2861e78272 | ||
|
|
494d598f86 | ||
|
|
c6208f5897 | ||
|
|
8f33cc6094 | ||
|
|
2ec9cf394a | ||
|
|
770e433a9e | ||
|
|
c6af90553e | ||
|
|
7989d1fcc8 | ||
|
|
611008fa86 | ||
|
|
bf628887f1 | ||
|
|
330dd8472f | ||
|
|
b530f1cf42 | ||
|
|
8ee551c99a | ||
|
|
148014e709 | ||
|
|
f2645c5df8 | ||
|
|
9ac81370fa | ||
|
|
93cafa8fe9 | ||
|
|
1627e96677 | ||
|
|
7e90d36120 | ||
|
|
9e50fdc01f | ||
|
|
2c0b36a967 | ||
|
|
83397dc811 | ||
|
|
640b566412 | ||
|
|
da24673a9f | ||
|
|
bded2585a4 | ||
|
|
5384563c57 | ||
|
|
01a62e2bc0 | ||
|
|
02fdb070ad | ||
|
|
278f4adbd2 | ||
|
|
d00076a7c1 | ||
|
|
4672b8ba2b | ||
|
|
56e669e2e4 | ||
|
|
a1b5a6f6e3 | ||
|
|
a4fba501a6 | ||
|
|
89789bd848 | ||
|
|
8f6b189dae | ||
|
|
69522dd5b7 | ||
|
|
eddc6fd7e1 | ||
|
|
7ceb3122fc | ||
|
|
62781dc15e | ||
|
|
3f97ef1288 | ||
|
|
408c9f69a8 | ||
|
|
5df07f114e | ||
|
|
541e9fc3f5 | ||
|
|
b6f3b7add9 |
79
CHANGELOG.md
79
CHANGELOG.md
@@ -1,5 +1,84 @@
|
|||||||
## Changelog
|
## Changelog
|
||||||
|
|
||||||
|
### 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
|
### CUDA 12.9
|
||||||
* Updated toolchain for cross-compilation for Tegra Linux platforms.
|
* Updated toolchain for cross-compilation for Tegra Linux platforms.
|
||||||
* Added `run_tests.py` utility to exercise all samples. See README.md for details
|
* Added `run_tests.py` utility to exercise all samples. See README.md for details
|
||||||
|
|||||||
@@ -1,7 +1,5 @@
|
|||||||
cmake_minimum_required(VERSION 3.20)
|
cmake_minimum_required(VERSION 3.20)
|
||||||
|
|
||||||
list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/../../../cmake/Modules")
|
|
||||||
|
|
||||||
project(cuda-samples LANGUAGES C CXX CUDA)
|
project(cuda-samples LANGUAGES C CXX CUDA)
|
||||||
|
|
||||||
find_package(CUDAToolkit REQUIRED)
|
find_package(CUDAToolkit REQUIRED)
|
||||||
@@ -14,7 +12,7 @@ set(CMAKE_CXX_STANDARD_REQUIRED ON)
|
|||||||
set(CMAKE_CUDA_STANDARD 17)
|
set(CMAKE_CUDA_STANDARD 17)
|
||||||
set(CMAKE_CUDA_STANDARD_REQUIRED ON)
|
set(CMAKE_CUDA_STANDARD_REQUIRED ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -138,26 +138,29 @@ 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>";
|
||||||
@@ -670,7 +673,10 @@ inline int _ConvertSMVer2Cores(int major, int minor) {
|
|||||||
{0x90, 128},
|
{0x90, 128},
|
||||||
{0xa0, 128},
|
{0xa0, 128},
|
||||||
{0xa1, 128},
|
{0xa1, 128},
|
||||||
|
{0xa3, 128},
|
||||||
|
{0xb0, 128},
|
||||||
{0xc0, 128},
|
{0xc0, 128},
|
||||||
|
{0xc1, 128},
|
||||||
{-1, -1}};
|
{-1, -1}};
|
||||||
|
|
||||||
int index = 0;
|
int index = 0;
|
||||||
@@ -722,7 +728,10 @@ inline const char* _ConvertSMVer2ArchName(int major, int minor) {
|
|||||||
{0x90, "Hopper"},
|
{0x90, "Hopper"},
|
||||||
{0xa0, "Blackwell"},
|
{0xa0, "Blackwell"},
|
||||||
{0xa1, "Blackwell"},
|
{0xa1, "Blackwell"},
|
||||||
|
{0xa3, "Blackwell"},
|
||||||
|
{0xb0, "Blackwell"},
|
||||||
{0xc0, "Blackwell"},
|
{0xc0, "Blackwell"},
|
||||||
|
{0xc1, "Blackwell"},
|
||||||
{-1, "Graphics Device"}};
|
{-1, "Graphics Device"}};
|
||||||
|
|
||||||
int index = 0;
|
int index = 0;
|
||||||
|
|||||||
@@ -118,7 +118,10 @@ inline int _ConvertSMVer2CoresDRV(int major, int minor) {
|
|||||||
{0x90, 128},
|
{0x90, 128},
|
||||||
{0xa0, 128},
|
{0xa0, 128},
|
||||||
{0xa1, 128},
|
{0xa1, 128},
|
||||||
|
{0xa3, 128},
|
||||||
|
{0xb0, 128},
|
||||||
{0xc0, 128},
|
{0xc0, 128},
|
||||||
|
{0xc1, 128},
|
||||||
{-1, -1}};
|
{-1, -1}};
|
||||||
|
|
||||||
int index = 0;
|
int index = 0;
|
||||||
|
|||||||
@@ -185,25 +185,30 @@ int ipcCreateSocket(ipcHandle *&handle, const char *name,
|
|||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
unlink(name);
|
char path_name[50];
|
||||||
bzero(&servaddr, sizeof(servaddr));
|
|
||||||
|
// Create unique name for the socket with path if SOCK_FOLDER is set.
|
||||||
|
sprintf(path_name, "%s%u", SOCK_FOLDER, getpid());
|
||||||
|
|
||||||
|
unlink(path_name);
|
||||||
|
memset(&servaddr, 0, sizeof(servaddr));
|
||||||
servaddr.sun_family = AF_UNIX;
|
servaddr.sun_family = AF_UNIX;
|
||||||
|
|
||||||
size_t len = strlen(name);
|
size_t len = strlen(path_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, name, len);
|
strncpy(servaddr.sun_path, 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(name) + 1];
|
handle->socketName = new char[strlen(path_name) + 1];
|
||||||
strcpy(handle->socketName, name);
|
strcpy(handle->socketName, path_name);
|
||||||
handle->socket = server_fd;
|
handle->socket = server_fd;
|
||||||
return 0;
|
return 0;
|
||||||
}
|
}
|
||||||
@@ -220,12 +225,12 @@ int ipcOpenSocket(ipcHandle *&handle) {
|
|||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
bzero(&cliaddr, sizeof(cliaddr));
|
memset(&cliaddr, 0, sizeof(cliaddr));
|
||||||
cliaddr.sun_family = AF_UNIX;
|
cliaddr.sun_family = AF_UNIX;
|
||||||
char temp[10];
|
char temp[50];
|
||||||
|
|
||||||
// Create unique name for the socket.
|
// Create unique name for the socket with path if SOCK_FOLDER is set.
|
||||||
sprintf(temp, "%u", getpid());
|
sprintf(temp, "%s%u", SOCK_FOLDER, 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) {
|
||||||
@@ -326,7 +331,7 @@ int ipcSendDataToServer(ipcHandle *handle, const char *serverName,
|
|||||||
ssize_t sendResult;
|
ssize_t sendResult;
|
||||||
struct sockaddr_un serveraddr;
|
struct sockaddr_un serveraddr;
|
||||||
|
|
||||||
bzero(&serveraddr, sizeof(serveraddr));
|
memset(&serveraddr, 0, 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);
|
||||||
|
|
||||||
@@ -359,10 +364,10 @@ int ipcSendShareableHandle(ipcHandle *handle,
|
|||||||
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
|
||||||
bzero(&cliaddr, sizeof(cliaddr));
|
memset(&cliaddr, 0, sizeof(cliaddr));
|
||||||
cliaddr.sun_family = AF_UNIX;
|
cliaddr.sun_family = AF_UNIX;
|
||||||
char temp[10];
|
char temp[20];
|
||||||
sprintf(temp, "%u", process);
|
sprintf(temp, "%s%u", SOCK_FOLDER, process);
|
||||||
strcpy(cliaddr.sun_path, temp);
|
strcpy(cliaddr.sun_path, temp);
|
||||||
len = sizeof(cliaddr);
|
len = sizeof(cliaddr);
|
||||||
|
|
||||||
|
|||||||
@@ -54,6 +54,13 @@
|
|||||||
#endif
|
#endif
|
||||||
#include <vector>
|
#include <vector>
|
||||||
|
|
||||||
|
// Define "/tmp" as socket creating folder for QNX
|
||||||
|
#if defined(__QNX__)
|
||||||
|
#define SOCK_FOLDER "/tmp/"
|
||||||
|
#else
|
||||||
|
#define SOCK_FOLDER ""
|
||||||
|
#endif
|
||||||
|
|
||||||
typedef struct sharedMemoryInfo_st {
|
typedef struct sharedMemoryInfo_st {
|
||||||
void *addr;
|
void *addr;
|
||||||
size_t size;
|
size_t size;
|
||||||
|
|||||||
@@ -187,6 +187,7 @@ 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);
|
||||||
@@ -200,7 +201,7 @@ 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, 0, cuDevice));
|
checkCudaErrors(cuCtxCreate(&context, &ctxCreateParams, 0, cuDevice));
|
||||||
|
|
||||||
checkCudaErrors(cuModuleLoadData(&module, cubin));
|
checkCudaErrors(cuModuleLoadData(&module, cubin));
|
||||||
free(cubin);
|
free(cubin);
|
||||||
|
|||||||
82
README.md
82
README.md
@@ -1,6 +1,6 @@
|
|||||||
# CUDA Samples
|
# CUDA Samples
|
||||||
|
|
||||||
Samples for CUDA Developers which demonstrates features in CUDA Toolkit. This version supports [CUDA Toolkit 12.9](https://developer.nvidia.com/cuda-downloads).
|
Samples for CUDA Developers which demonstrates features in CUDA Toolkit. This version supports [CUDA Toolkit 13.0](https://developer.nvidia.com/cuda-downloads).
|
||||||
|
|
||||||
## Release Notes
|
## Release Notes
|
||||||
|
|
||||||
@@ -115,29 +115,85 @@ make -j$(nproc)
|
|||||||
```
|
```
|
||||||
Transfer the built binaries to the Tegra device and execute them there.
|
Transfer the built binaries to the Tegra device and execute them there.
|
||||||
|
|
||||||
### Building for Automotive Linux Platforms
|
|
||||||
|
|
||||||
These platforms require additional information to be passed to CMake on the command line to ensure proper resolution of all necessary include and library files.
|
### Cross Building for Automotive Linux Platforms from the DriveOS Docker containers
|
||||||
Instead of being in the default location, `/usr/local/cuda/include` or `/usr/local/cuda/lib64`, you must point to architecture-specific paths:
|
|
||||||
|
|
||||||
`/usr/local/cuda/<ARCH>/targets/aarch64-linux/lib`
|
To build CUDA samples to the target platform from the DriveOS Docker containers, use the following instructions.
|
||||||
and
|
|
||||||
`/usr/local/cuda/<ARCH>/include`
|
|
||||||
|
|
||||||
An example build might look like this:
|
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-13.0/thor/lib64/ \
|
||||||
|
-DCMAKE_INCLUDE_PATH=/drive/temp/usr/local/cuda-13.0/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 Samples/5_Domain_Specific/CMakeList.txt
|
||||||
|
# add_subdirectory(simpleGL)
|
||||||
|
# add_subdirectory(simpleVulkan)
|
||||||
|
# add_subdirectory(simpleVulkanMMAP)
|
||||||
|
|
||||||
|
# In Samples/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
|
$ mkdir build
|
||||||
$ cd build
|
$ cd build
|
||||||
|
|
||||||
$ cmake -DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc -DCMAKE_LIBRARY_PATH=/usr/local/cuda/orin/lib64/ -DCMAKE_INCLUDE_PATH=/usr/local/cuda/orin/include -DBUILD_TEGRA=True ..
|
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/
|
||||||
```
|
```
|
||||||
|
|
||||||
### QNX
|
### Forward Compatibility
|
||||||
|
|
||||||
Note that in the current branch sample cross-compilation for QNX is not fully validated. This placeholder will be updated in the
|
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:
|
||||||
near future with QNX cross-compilation instructions. In the meantime, if you want to cross-compile for QNX please check out one
|
|
||||||
of the previous tags prior to the CMake build system transition in 12.8.
|
```
|
||||||
|
cmake -DCMAKE_PREFIX_PATH=/usr/local/cuda/lib64/stubs/ ..
|
||||||
|
```
|
||||||
|
|
||||||
## Running All Samples as Tests
|
## Running All Samples as Tests
|
||||||
|
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -19,6 +19,12 @@ endif()
|
|||||||
# Include directories and libraries
|
# Include directories and libraries
|
||||||
include_directories(../../../Common)
|
include_directories(../../../Common)
|
||||||
|
|
||||||
|
# This sample is not supported on QNX
|
||||||
|
if(CMAKE_SYSTEM_NAME STREQUAL "QNX")
|
||||||
|
message(STATUS "Will not build sample ${PROJECT_NAME} - not supported on QNX")
|
||||||
|
return()
|
||||||
|
endif()
|
||||||
|
|
||||||
# Source file
|
# Source file
|
||||||
if(CMAKE_GENERATOR MATCHES "Visual Studio")
|
if(CMAKE_GENERATOR MATCHES "Visual Studio")
|
||||||
find_package(OpenMP REQUIRED C CXX)
|
find_package(OpenMP REQUIRED C CXX)
|
||||||
|
|||||||
@@ -247,7 +247,9 @@ int main(int argc, char **argv)
|
|||||||
exit(EXIT_WAIVED);
|
exit(EXIT_WAIVED);
|
||||||
}
|
}
|
||||||
|
|
||||||
if (device_prop.computeMode == cudaComputeModeProhibited) {
|
int computeMode;
|
||||||
|
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, dev_id));
|
||||||
|
if (computeMode == cudaComputeModeProhibited) {
|
||||||
// This sample requires being run with a default or process exclusive mode
|
// This sample requires being run with a default or process exclusive mode
|
||||||
fprintf(stderr,
|
fprintf(stderr,
|
||||||
"This sample requires a device in either default or process "
|
"This sample requires a device in either default or process "
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
15
Samples/0_Introduction/clock/.vscode/tasks.json
vendored
15
Samples/0_Introduction/clock/.vscode/tasks.json
vendored
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
else()
|
else()
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -46,6 +46,14 @@ foreach(arch ${CMAKE_CUDA_ARCHITECTURES})
|
|||||||
list(APPEND GENCODE_FLAGS "-gencode=arch=compute_${arch},code=sm_${arch}")
|
list(APPEND GENCODE_FLAGS "-gencode=arch=compute_${arch},code=sm_${arch}")
|
||||||
endforeach()
|
endforeach()
|
||||||
|
|
||||||
|
if(CMAKE_SYSTEM_NAME STREQUAL "QNX")
|
||||||
|
set(INCLUDES_LIST)
|
||||||
|
foreach(dir ${CUDAToolkit_INCLUDE_DIRS})
|
||||||
|
list(APPEND INCLUDES_LIST "-I${dir}")
|
||||||
|
endforeach()
|
||||||
|
string(JOIN " " INCLUDES "${INCLUDES_LIST}")
|
||||||
|
endif()
|
||||||
|
|
||||||
add_custom_command(
|
add_custom_command(
|
||||||
OUTPUT ${CUDA_FATBIN_FILE}
|
OUTPUT ${CUDA_FATBIN_FILE}
|
||||||
COMMAND ${CMAKE_CUDA_COMPILER} ${INCLUDES} ${ALL_CCFLAGS} -Wno-deprecated-gpu-targets ${GENCODE_FLAGS} -o ${CUDA_FATBIN_FILE} -fatbin ${CUDA_KERNEL_SOURCE}
|
COMMAND ${CMAKE_CUDA_COMPILER} ${INCLUDES} ${ALL_CCFLAGS} -Wno-deprecated-gpu-targets ${GENCODE_FLAGS} -o ${CUDA_FATBIN_FILE} -fatbin ${CUDA_KERNEL_SOURCE}
|
||||||
|
|||||||
@@ -268,9 +268,10 @@ void randomInit(float *data, int size)
|
|||||||
|
|
||||||
static int initCUDA(int argc, char **argv, CUfunction *pMatrixMul, int *blk_size)
|
static int initCUDA(int argc, char **argv, CUfunction *pMatrixMul, int *blk_size)
|
||||||
{
|
{
|
||||||
CUfunction cuFunction = 0;
|
CUfunction cuFunction = 0;
|
||||||
int major = 0, minor = 0;
|
int major = 0, minor = 0;
|
||||||
char deviceName[100];
|
char deviceName[100];
|
||||||
|
CUctxCreateParams ctxCreateParams = {};
|
||||||
|
|
||||||
cuDevice = findCudaDeviceDRV(argc, (const char **)argv);
|
cuDevice = findCudaDeviceDRV(argc, (const char **)argv);
|
||||||
|
|
||||||
@@ -283,7 +284,7 @@ static int initCUDA(int argc, char **argv, CUfunction *pMatrixMul, int *blk_size
|
|||||||
checkCudaErrors(cuDeviceTotalMem(&totalGlobalMem, cuDevice));
|
checkCudaErrors(cuDeviceTotalMem(&totalGlobalMem, cuDevice));
|
||||||
printf(" Total amount of global memory: %llu bytes\n", (long long unsigned int)totalGlobalMem);
|
printf(" Total amount of global memory: %llu bytes\n", (long long unsigned int)totalGlobalMem);
|
||||||
|
|
||||||
checkCudaErrors(cuCtxCreate(&cuContext, 0, cuDevice));
|
checkCudaErrors(cuCtxCreate(&cuContext, &ctxCreateParams, 0, cuDevice));
|
||||||
|
|
||||||
// first search for the module path before we load the results
|
// first search for the module path before we load the results
|
||||||
std::string module_path;
|
std::string module_path;
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -31,12 +31,17 @@ set_target_properties(matrixMulDynlinkJIT PROPERTIES
|
|||||||
CUDA_SEPARABLE_COMPILATION ON
|
CUDA_SEPARABLE_COMPILATION ON
|
||||||
POSITION_INDEPENDENT_CODE OFF
|
POSITION_INDEPENDENT_CODE OFF
|
||||||
)
|
)
|
||||||
set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -no-pie")
|
|
||||||
|
# Only add -no-pie for GCC or Clang
|
||||||
|
if (CMAKE_CXX_COMPILER_ID STREQUAL "GNU" OR CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
|
||||||
|
set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -no-pie")
|
||||||
|
endif()
|
||||||
|
|
||||||
target_link_libraries(matrixMulDynlinkJIT PUBLIC
|
target_link_libraries(matrixMulDynlinkJIT PUBLIC
|
||||||
CUDA::cudart
|
CUDA::cudart
|
||||||
CUDA::cuda_driver
|
CUDA::cuda_driver
|
||||||
)
|
)
|
||||||
|
|
||||||
if(UNIX)
|
if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
|
||||||
target_link_libraries(matrixMulDynlinkJIT PUBLIC dl)
|
target_link_libraries(matrixMulDynlinkJIT PUBLIC dl)
|
||||||
endif()
|
endif()
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -32,23 +32,26 @@ target_link_libraries(matrixMul_nvrtc PRIVATE
|
|||||||
CUDA::cuda_driver
|
CUDA::cuda_driver
|
||||||
)
|
)
|
||||||
|
|
||||||
|
# The primary directory of CUDAToolkit_INCLUDE_DIRS is the CUDA Toolkit's include directory for finding the header files.
|
||||||
|
list(GET CUDAToolkit_INCLUDE_DIRS 0 CUDA_INCLUDE_DIR)
|
||||||
|
|
||||||
# Copy clock_kernel.cu to the output directory
|
# Copy clock_kernel.cu to the output directory
|
||||||
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
||||||
COMMAND ${CMAKE_COMMAND} -E copy_if_different
|
COMMAND ${CMAKE_COMMAND} -E copy_if_different
|
||||||
${CMAKE_CURRENT_SOURCE_DIR}/matrixMul_kernel.cu ${CUDAToolkit_BIN_DIR}/../include/cooperative_groups.h ${CMAKE_CURRENT_BINARY_DIR}
|
${CMAKE_CURRENT_SOURCE_DIR}/matrixMul_kernel.cu ${CUDA_INCLUDE_DIR}/cooperative_groups.h ${CMAKE_CURRENT_BINARY_DIR}
|
||||||
)
|
)
|
||||||
|
|
||||||
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
||||||
COMMAND ${CMAKE_COMMAND} -E copy_directory
|
COMMAND ${CMAKE_COMMAND} -E copy_directory
|
||||||
${CUDAToolkit_BIN_DIR}/../include/cooperative_groups ${CMAKE_CURRENT_BINARY_DIR}/cooperative_groups
|
${CUDA_INCLUDE_DIR}/cooperative_groups ${CMAKE_CURRENT_BINARY_DIR}/cooperative_groups
|
||||||
)
|
)
|
||||||
|
|
||||||
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
||||||
COMMAND ${CMAKE_COMMAND} -E copy_directory
|
COMMAND ${CMAKE_COMMAND} -E copy_directory
|
||||||
${CUDAToolkit_BIN_DIR}/../include/nv ${CMAKE_CURRENT_BINARY_DIR}/nv
|
${CUDA_INCLUDE_DIR}/cccl/nv ${CMAKE_CURRENT_BINARY_DIR}/nv
|
||||||
)
|
)
|
||||||
|
|
||||||
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
add_custom_command(TARGET matrixMul_nvrtc POST_BUILD
|
||||||
COMMAND ${CMAKE_COMMAND} -E copy_directory
|
COMMAND ${CMAKE_COMMAND} -E copy_directory
|
||||||
${CUDAToolkit_BIN_DIR}/../include/cuda ${CMAKE_CURRENT_BINARY_DIR}/cuda
|
${CUDA_INCLUDE_DIR}/cccl/cuda ${CMAKE_CURRENT_BINARY_DIR}/cuda
|
||||||
)
|
)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
|
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
@@ -20,6 +20,12 @@ endif()
|
|||||||
# Include directories and libraries
|
# Include directories and libraries
|
||||||
include_directories(../../../Common)
|
include_directories(../../../Common)
|
||||||
|
|
||||||
|
# This sample is not supported on QNX
|
||||||
|
if(CMAKE_SYSTEM_NAME STREQUAL "QNX")
|
||||||
|
message(STATUS "simpleAWBarrier is not supported on QNX")
|
||||||
|
return()
|
||||||
|
endif()
|
||||||
|
|
||||||
# Source file
|
# Source file
|
||||||
# Add target for simpleAWBarrier
|
# Add target for simpleAWBarrier
|
||||||
add_executable(simpleAWBarrier simpleAWBarrier.cu)
|
add_executable(simpleAWBarrier simpleAWBarrier.cu)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
|
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -48,6 +48,14 @@ foreach(arch ${CMAKE_CUDA_ARCHITECTURES})
|
|||||||
list(APPEND GENCODE_FLAGS "-gencode=arch=compute_${arch},code=sm_${arch}")
|
list(APPEND GENCODE_FLAGS "-gencode=arch=compute_${arch},code=sm_${arch}")
|
||||||
endforeach()
|
endforeach()
|
||||||
|
|
||||||
|
if(CMAKE_SYSTEM_NAME STREQUAL "QNX")
|
||||||
|
set(INCLUDES_LIST)
|
||||||
|
foreach(dir ${CUDAToolkit_INCLUDE_DIRS})
|
||||||
|
list(APPEND INCLUDES_LIST "-I${dir}")
|
||||||
|
endforeach()
|
||||||
|
string(JOIN " " INCLUDES "${INCLUDES_LIST}")
|
||||||
|
endif()
|
||||||
|
|
||||||
add_custom_command(
|
add_custom_command(
|
||||||
OUTPUT ${CUDA_FATBIN_FILE}
|
OUTPUT ${CUDA_FATBIN_FILE}
|
||||||
COMMAND ${CMAKE_CUDA_COMPILER} ${INCLUDES} ${ALL_CCFLAGS} -Wno-deprecated-gpu-targets ${GENCODE_FLAGS} -o ${CUDA_FATBIN_FILE} -fatbin ${CUDA_KERNEL_SOURCE}
|
COMMAND ${CMAKE_CUDA_COMPILER} ${INCLUDES} ${ALL_CCFLAGS} -Wno-deprecated-gpu-targets ${GENCODE_FLAGS} -o ${CUDA_FATBIN_FILE} -fatbin ${CUDA_KERNEL_SOURCE}
|
||||||
|
|||||||
@@ -80,19 +80,20 @@ static void check(CUresult result, char const *const func, const char *const fil
|
|||||||
int main(int argc, char **argv)
|
int main(int argc, char **argv)
|
||||||
{
|
{
|
||||||
printf("simpleDrvRuntime..\n");
|
printf("simpleDrvRuntime..\n");
|
||||||
int N = 50000, devID = 0;
|
int N = 50000, devID = 0;
|
||||||
size_t size = N * sizeof(float);
|
size_t size = N * sizeof(float);
|
||||||
CUdevice cuDevice;
|
CUdevice cuDevice;
|
||||||
CUfunction vecAdd_kernel;
|
CUfunction vecAdd_kernel;
|
||||||
CUmodule cuModule = 0;
|
CUmodule cuModule = 0;
|
||||||
CUcontext cuContext;
|
CUcontext cuContext;
|
||||||
|
CUctxCreateParams ctxCreateParams = {};
|
||||||
|
|
||||||
// Initialize
|
// Initialize
|
||||||
checkCudaDrvErrors(cuInit(0));
|
checkCudaDrvErrors(cuInit(0));
|
||||||
|
|
||||||
cuDevice = findCudaDevice(argc, (const char **)argv);
|
cuDevice = findCudaDevice(argc, (const char **)argv);
|
||||||
// Create context
|
// Create context
|
||||||
checkCudaDrvErrors(cuCtxCreate(&cuContext, 0, cuDevice));
|
checkCudaDrvErrors(cuCtxCreate(&cuContext, &ctxCreateParams, 0, cuDevice));
|
||||||
|
|
||||||
// first search for the module path before we load the results
|
// first search for the module path before we load the results
|
||||||
string module_path;
|
string module_path;
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -127,6 +127,10 @@ int main(int argc, char **argv)
|
|||||||
checkCudaErrors(cudaGetDevice(&cuda_device));
|
checkCudaErrors(cudaGetDevice(&cuda_device));
|
||||||
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, cuda_device));
|
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, cuda_device));
|
||||||
|
|
||||||
|
// Get device clock rate
|
||||||
|
int clockRate;
|
||||||
|
checkCudaErrors(cudaDeviceGetAttribute(&clockRate, cudaDevAttrClockRate, cuda_device));
|
||||||
|
|
||||||
// HyperQ is available in devices of Compute Capability 3.5 and higher
|
// HyperQ is available in devices of Compute Capability 3.5 and higher
|
||||||
if (deviceProp.major < 3 || (deviceProp.major == 3 && deviceProp.minor < 5)) {
|
if (deviceProp.major < 3 || (deviceProp.major == 3 && deviceProp.minor < 5)) {
|
||||||
if (deviceProp.concurrentKernels == 0) {
|
if (deviceProp.concurrentKernels == 0) {
|
||||||
@@ -170,9 +174,9 @@ int main(int argc, char **argv)
|
|||||||
#if defined(__arm__) || defined(__aarch64__)
|
#if defined(__arm__) || defined(__aarch64__)
|
||||||
// the kernel takes more time than the channel reset time on arm archs, so to
|
// the kernel takes more time than the channel reset time on arm archs, so to
|
||||||
// prevent hangs reduce time_clocks.
|
// prevent hangs reduce time_clocks.
|
||||||
clock_t time_clocks = (clock_t)(kernel_time * (deviceProp.clockRate / 100));
|
clock_t time_clocks = (clock_t)(kernel_time * (clockRate / 100));
|
||||||
#else
|
#else
|
||||||
clock_t time_clocks = (clock_t)(kernel_time * deviceProp.clockRate);
|
clock_t time_clocks = (clock_t)(kernel_time * clockRate);
|
||||||
#endif
|
#endif
|
||||||
clock_t total_clocks = 0;
|
clock_t total_clocks = 0;
|
||||||
|
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -247,7 +247,9 @@ static void parentProcess(char *app)
|
|||||||
}
|
}
|
||||||
// This sample requires two processes accessing each device, so we need
|
// This sample requires two processes accessing each device, so we need
|
||||||
// to ensure exclusive or prohibited mode is not set
|
// to ensure exclusive or prohibited mode is not set
|
||||||
if (prop.computeMode != cudaComputeModeDefault) {
|
int computeMode;
|
||||||
|
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, i));
|
||||||
|
if (computeMode != cudaComputeModeDefault) {
|
||||||
printf("Device %d is in an unsupported compute mode for this sample\n", i);
|
printf("Device %d is in an unsupported compute mode for this sample\n", i);
|
||||||
continue;
|
continue;
|
||||||
}
|
}
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -37,6 +37,7 @@ target_compile_features(simpleMPI PRIVATE cxx_std_17 cuda_std_17)
|
|||||||
)
|
)
|
||||||
|
|
||||||
target_link_libraries(simpleMPI PUBLIC
|
target_link_libraries(simpleMPI PUBLIC
|
||||||
|
${MPI_C_LIBRARIES}
|
||||||
${MPI_CXX_LIBRARIES}
|
${MPI_CXX_LIBRARIES}
|
||||||
)
|
)
|
||||||
|
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -218,9 +218,11 @@ int main(int argc, char *argv[])
|
|||||||
|
|
||||||
printf("\n");
|
printf("\n");
|
||||||
printf("Relevant properties of this CUDA device\n");
|
printf("Relevant properties of this CUDA device\n");
|
||||||
|
int canOverlap;
|
||||||
|
checkCudaErrors(cudaDeviceGetAttribute(&canOverlap, cudaDevAttrGpuOverlap, cuda_device));
|
||||||
printf("(%s) Can overlap one CPU<>GPU data transfer with GPU kernel execution "
|
printf("(%s) Can overlap one CPU<>GPU data transfer with GPU kernel execution "
|
||||||
"(device property \"deviceOverlap\")\n",
|
"(device property \"cudaDevAttrGpuOverlap\")\n",
|
||||||
deviceProp.deviceOverlap ? "X" : " ");
|
canOverlap ? "X" : " ");
|
||||||
// printf("(%s) Can execute several GPU kernels simultaneously (compute
|
// printf("(%s) Can execute several GPU kernels simultaneously (compute
|
||||||
// capability >= 2.0)\n", deviceProp.major >= 2 ? "X": " ");
|
// capability >= 2.0)\n", deviceProp.major >= 2 ? "X": " ");
|
||||||
printf("(%s) Can overlap two CPU<>GPU data transfers with GPU kernel execution\n"
|
printf("(%s) Can overlap two CPU<>GPU data transfers with GPU kernel execution\n"
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
@@ -45,6 +45,14 @@ foreach(arch ${CMAKE_CUDA_ARCHITECTURES})
|
|||||||
list(APPEND GENCODE_FLAGS "-gencode=arch=compute_${arch},code=sm_${arch}")
|
list(APPEND GENCODE_FLAGS "-gencode=arch=compute_${arch},code=sm_${arch}")
|
||||||
endforeach()
|
endforeach()
|
||||||
|
|
||||||
|
if(CMAKE_SYSTEM_NAME STREQUAL "QNX")
|
||||||
|
set(INCLUDES_LIST)
|
||||||
|
foreach(dir ${CUDAToolkit_INCLUDE_DIRS})
|
||||||
|
list(APPEND INCLUDES_LIST "-I${dir}")
|
||||||
|
endforeach()
|
||||||
|
string(JOIN " " INCLUDES "${INCLUDES_LIST}")
|
||||||
|
endif()
|
||||||
|
|
||||||
add_custom_command(
|
add_custom_command(
|
||||||
OUTPUT ${CUDA_FATBIN_FILE}
|
OUTPUT ${CUDA_FATBIN_FILE}
|
||||||
COMMAND ${CMAKE_CUDA_COMPILER} ${INCLUDES} ${ALL_CCFLAGS} -Wno-deprecated-gpu-targets ${GENCODE_FLAGS} -o ${CUDA_FATBIN_FILE} -fatbin ${CUDA_KERNEL_SOURCE}
|
COMMAND ${CMAKE_CUDA_COMPILER} ${INCLUDES} ${ALL_CCFLAGS} -Wno-deprecated-gpu-targets ${GENCODE_FLAGS} -o ${CUDA_FATBIN_FILE} -fatbin ${CUDA_KERNEL_SOURCE}
|
||||||
|
|||||||
@@ -309,10 +309,11 @@ void runTest(int argc, char **argv)
|
|||||||
////////////////////////////////////////////////////////////////////////////////
|
////////////////////////////////////////////////////////////////////////////////
|
||||||
static CUresult initCUDA(int argc, char **argv, CUfunction *transform)
|
static CUresult initCUDA(int argc, char **argv, CUfunction *transform)
|
||||||
{
|
{
|
||||||
CUfunction cuFunction = 0;
|
CUfunction cuFunction = 0;
|
||||||
int major = 0, minor = 0, devID = 0;
|
int major = 0, minor = 0, devID = 0;
|
||||||
char deviceName[100];
|
char deviceName[100];
|
||||||
string module_path;
|
string module_path;
|
||||||
|
CUctxCreateParams ctxCreateParams = {};
|
||||||
|
|
||||||
cuDevice = findCudaDeviceDRV(argc, (const char **)argv);
|
cuDevice = findCudaDeviceDRV(argc, (const char **)argv);
|
||||||
|
|
||||||
@@ -322,7 +323,7 @@ static CUresult initCUDA(int argc, char **argv, CUfunction *transform)
|
|||||||
checkCudaErrors(cuDeviceGetName(deviceName, sizeof(deviceName), cuDevice));
|
checkCudaErrors(cuDeviceGetName(deviceName, sizeof(deviceName), cuDevice));
|
||||||
printf("> GPU Device has SM %d.%d compute capability\n", major, minor);
|
printf("> GPU Device has SM %d.%d compute capability\n", major, minor);
|
||||||
|
|
||||||
checkCudaErrors(cuCtxCreate(&cuContext, 0, cuDevice));
|
checkCudaErrors(cuCtxCreate(&cuContext, &ctxCreateParams, 0, cuDevice));
|
||||||
|
|
||||||
// first search for the module_path before we try to load the results
|
// first search for the module_path before we try to load the results
|
||||||
std::ostringstream fatbin;
|
std::ostringstream fatbin;
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
else()
|
else()
|
||||||
|
|||||||
@@ -287,7 +287,9 @@ int main(int argc, char **argv)
|
|||||||
exit(EXIT_WAIVED);
|
exit(EXIT_WAIVED);
|
||||||
}
|
}
|
||||||
|
|
||||||
if (device_prop.computeMode == cudaComputeModeProhibited) {
|
int computeMode;
|
||||||
|
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, dev_id));
|
||||||
|
if (computeMode == cudaComputeModeProhibited) {
|
||||||
// This sample requires being run with a default or process exclusive mode
|
// This sample requires being run with a default or process exclusive mode
|
||||||
fprintf(stderr,
|
fprintf(stderr,
|
||||||
"This sample requires a device in either default or process "
|
"This sample requires a device in either default or process "
|
||||||
|
|||||||
@@ -1,15 +0,0 @@
|
|||||||
{
|
|
||||||
"version": "2.0.0",
|
|
||||||
"tasks": [
|
|
||||||
{
|
|
||||||
"label": "sample",
|
|
||||||
"type": "shell",
|
|
||||||
"command": "make dbg=1",
|
|
||||||
"problemMatcher": ["$nvcc"],
|
|
||||||
"group": {
|
|
||||||
"kind": "build",
|
|
||||||
"isDefault": true
|
|
||||||
}
|
|
||||||
}
|
|
||||||
]
|
|
||||||
}
|
|
||||||
@@ -8,7 +8,7 @@ find_package(CUDAToolkit REQUIRED)
|
|||||||
|
|
||||||
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
|
||||||
|
|
||||||
set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120)
|
set(CMAKE_CUDA_ARCHITECTURES 75 80 86 87 89 90 100 110 120)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
|
||||||
if(ENABLE_CUDA_DEBUG)
|
if(ENABLE_CUDA_DEBUG)
|
||||||
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -G") # enable cuda-gdb (may significantly affect performance on some targets)
|
||||||
|
|||||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user