Add and update samples for cuda 11.0 support

2026-10-23 21:28:17 +08:00 · 2020-05-18 22:22:06 +05:30
parent 6be514679b
commit 4f6e02970b
410 changed files with 21920 additions and 3864 deletions
--- a/Samples/cannyEdgeDetectorNPP/Makefile
+++ b/Samples/cannyEdgeDetectorNPP/Makefile
@@ -72,9 +72,9 @@ endif
 # architecture
 HOST_ARCH   := $(shell uname -m)
 TARGET_ARCH ?= $(HOST_ARCH)
-ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 ppc64le armv7l))
+ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
    ifneq ($(TARGET_ARCH),$(HOST_ARCH))
-        ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 ppc64le))
+        ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
            TARGET_SIZE := 64
        else ifneq (,$(filter $(TARGET_ARCH),armv7l))
            TARGET_SIZE := 32
@@ -85,8 +85,17 @@ ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 ppc64le armv7l))
 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 $(CUDA_PATH)/targets/sbsa-linux))
+        HOST_ARCH := sbsa
+        TARGET_ARCH := sbsa
+    endif
+endif
+
 ifneq ($(TARGET_ARCH),$(HOST_ARCH))
-    ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-ppc64le))
+    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 $(TARGET_ARCH) is not supported!)
    endif
 endif
@@ -141,6 +150,8 @@ else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
        else ifeq ($(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
@@ -185,19 +196,27 @@ ifneq ($(TARGET_ARCH),$(HOST_ARCH))
                CCFLAGS += --sysroot=$(TARGET_FS)
            endif
            LDFLAGS += --sysroot=$(TARGET_FS)
-            LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L $(TARGET_FS)/lib
-            LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L $(TARGET_FS)/usr/lib
-            LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L $(TARGET_FS)/usr/lib/aarch64-linux-gnu
+            LDFLAGS += -rpath-link=$(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
-            CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu
+            CCFLAGS += -isystem=$(TARGET_FS)/usr/include  -I$(TARGET_FS)/usr/include
+            CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
+        endif
+    endif
+    ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
+        CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
+        LDFLAGS += -lsocket
+        LDFLAGS += -rpath=/usr/lib/aarch64-qnx-gnu -L/usr/lib/aarch64-qnx-gnu
+        ifneq ($(TARGET_FS),)
+            LDFLAGS += -rpath=$(TARGET_FS)/usr/lib -L $(TARGET_FS)/usr/lib
+            LDFLAGS += -rpath=$(TARGET_FS)/usr/libnvidia -L $(TARGET_FS)/usr/libnvidia
+        endif
+        ifdef TARGET_OVERRIDE # cuda toolkit targets override
+            NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
        endif
    endif
-endif
-
-ifeq ($(TARGET_OS),qnx)
-    CCFLAGS += -DWIN_INTERFACE_CUSTOM
-    LDFLAGS += -lsocket
 endif

 # Install directory of different arch
@@ -206,6 +225,8 @@ 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)
@@ -259,8 +280,8 @@ ifeq ($(GENCODE_FLAGS),)
 $(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))

 ifeq ($(SMS),)
-# Generate PTX code from SM 30
-GENCODE_FLAGS += -gencode arch=compute_30,code=compute_30
+# Generate PTX code from SM 35
+GENCODE_FLAGS += -gencode arch=compute_35,code=compute_35
 endif

 # Generate PTX code from the highest SM architecture in $(SMS) to guarantee forward-compatibility
--- a/Samples/cannyEdgeDetectorNPP/NsightEclipse.xml
+++ b/Samples/cannyEdgeDetectorNPP/NsightEclipse.xml
@@ -0,0 +1,83 @@
+<?xml version="1.0" encoding="UTF-8"?> 
+<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
+<entry>
+  <name>cannyEdgeDetectorNPP</name>
+  <description><![CDATA[An NPP CUDA Sample that demonstrates the recommended parameters to use with the nppiFilterCannyBorder_8u_C1R Canny Edge Detection image filter function. This function expects a single channel 8-bit grayscale input image. You can generate a grayscale image from a color image by first calling nppiColorToGray() or nppiRGBToGray(). The Canny Edge Detection function combines and improves on the techniques required to produce an edge detection image using multiple steps.]]></description>
+  <devicecompilation>whole</devicecompilation>
+  <fallback_min_ptx>true</fallback_min_ptx>
+  <includepaths>
+    <path>../../Common/UtilNPP</path>
+    <path>../../Common/FreeImage/include</path>
+    <path>./</path>
+    <path>../</path>
+    <path>../../common/inc</path>
+  </includepaths>
+  <keyconcepts>
+    <concept level="basic">Performance Strategies</concept>
+    <concept level="basic">Image Processing</concept>
+    <concept level="basic">NPP Library</concept>
+  </keyconcepts>
+  <keywords>
+    <keyword>CUDA</keyword>
+    <keyword>NPP</keyword>
+    <keyword>Image Processing</keyword>
+  </keywords>
+  <libraries>
+    <library>nppisu_static</library>
+    <library>nppif_static</library>
+    <library>nppc_static</library>
+    <library>culibos</library>
+    <library>freeimage</library>
+  </libraries>
+  <librarypaths>
+    <path>../../Common/FreeImage/lib/$(TARGET_OS)</path>
+    <path>../../Common/FreeImage/lib/$(TARGET_OS)/$(TARGET_ARCH)</path>
+  </librarypaths>
+  <nsight_eclipse>true</nsight_eclipse>
+  <primary_file>cannyEdgeDetectorNPP.cpp</primary_file>
+  <required_dependencies>
+    <dependency>FreeImage</dependency>
+    <dependency>NPP</dependency>
+  </required_dependencies>
+  <scopes>
+    <scope>1:CUDA Basic Topics</scope>
+    <scope>1:Performance Strategies</scope>
+    <scope>2:Image Processing</scope>
+    <scope>2:Computer Vision</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>
+  <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>Canny Edge Detector NPP</title>
+  <type>exe</type>
+</entry>
--- a/Samples/cannyEdgeDetectorNPP/README.md
+++ b/Samples/cannyEdgeDetectorNPP/README.md
@@ -10,11 +10,11 @@ Performance Strategies, Image Processing, NPP Library

 ## Supported SM Architectures

-[SM 3.0 ](https://developer.nvidia.com/cuda-gpus)  [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 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)

 ## Supported OSes

-Linux, Windows, MacOSX
+Linux, Windows

 ## Supported CPU Architecture

@@ -27,7 +27,7 @@ x86_64, ppc64le, armv7l

 ## Prerequisites

-Download and install the [CUDA Toolkit 10.2](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
+Download and install the [CUDA Toolkit 11.0](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
 Make sure the dependencies mentioned in [Dependencies]() section above are installed.

 ## Build and Run
@@ -67,29 +67,5 @@ The samples makefiles can take advantage of certain options:
    $ make HOST_COMPILER=g++
 ```

-### Mac
-The Mac samples are built using makefiles. To use the makefiles, change directory into the sample directory you wish to build, and run make:
-```
-$ cd <sample_dir>
-$ make
-```
-
-The samples makefiles can take advantage of certain options:
-
-*  **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="A B ..."
-    ```
-
-*  **HOST_COMPILER=<host_compiler>** - override the default clang host compiler. See the [Mac Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html#system-requirements) for a list of supported host compilers.
-    ```
-    $ make HOST_COMPILER=clang
-    ```
-
 ## References (for more details)

--- a/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP.cpp
+++ b/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP.cpp
@@ -47,6 +47,27 @@
 #include <helper_cuda.h>
 #include <helper_string.h>

+inline int cudaDeviceInit(int argc, const char **argv) {
+  int deviceCount;
+  checkCudaErrors(cudaGetDeviceCount(&deviceCount));
+
+  if (deviceCount == 0) {
+    std::cerr << "CUDA error: no devices supporting CUDA." << std::endl;
+    exit(EXIT_FAILURE);
+  }
+
+  int dev = findCudaDevice(argc, argv);
+
+  cudaDeviceProp deviceProp;
+  cudaGetDeviceProperties(&deviceProp, dev);
+  std::cerr << "cudaSetDevice GPU" << dev << " = " << deviceProp.name
+            << std::endl;
+
+  checkCudaErrors(cudaSetDevice(dev));
+
+  return dev;
+}
+
 bool printfNPPinfo(int argc, char *argv[]) {
  const NppLibraryVersion *libVer = nppGetLibVersion();

@@ -74,7 +95,7 @@ int main(int argc, char *argv[]) {
    std::string sFilename;
    char *filePath;

-    int dev = findCudaDevice(argc, (const char **)argv);
+    cudaDeviceInit(argc, (const char **)argv);

    if (printfNPPinfo(argc, argv) == false) {
      exit(EXIT_SUCCESS);
--- a/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2012.vcxproj
+++ b/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2012.vcxproj
@@ -62,7 +62,7 @@
      <OutputFile>$(OutDir)/cannyEdgeDetectorNPP.exe</OutputFile>
    </Link>
    <CudaCompile>
-      <CodeGeneration>compute_30,compute_30;</CodeGeneration>
+      <CodeGeneration>compute_35,compute_35;</CodeGeneration>
      <AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
      <Include>./;../../Common</Include>
      <Defines>WIN32</Defines>
--- a/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2013.vcxproj
+++ b/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2013.vcxproj
@@ -62,7 +62,7 @@
      <OutputFile>$(OutDir)/cannyEdgeDetectorNPP.exe</OutputFile>
    </Link>
    <CudaCompile>
-      <CodeGeneration>compute_30,compute_30;</CodeGeneration>
+      <CodeGeneration>compute_35,compute_35;</CodeGeneration>
      <AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
      <Include>./;../../Common</Include>
      <Defines>WIN32</Defines>
--- a/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2015.vcxproj
+++ b/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2015.vcxproj
@@ -62,7 +62,7 @@
      <OutputFile>$(OutDir)/cannyEdgeDetectorNPP.exe</OutputFile>
    </Link>
    <CudaCompile>
-      <CodeGeneration>compute_30,compute_30;</CodeGeneration>
+      <CodeGeneration>compute_35,compute_35;</CodeGeneration>
      <AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
      <Include>./;../../Common</Include>
      <Defines>WIN32</Defines>
--- a/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2017.vcxproj
+++ b/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2017.vcxproj
@@ -67,7 +67,7 @@
      <OutputFile>$(OutDir)/cannyEdgeDetectorNPP.exe</OutputFile>
    </Link>
    <CudaCompile>
-      <CodeGeneration>compute_30,compute_30;</CodeGeneration>
+      <CodeGeneration>compute_35,compute_35;</CodeGeneration>
      <AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
      <Include>./;../../Common</Include>
      <Defines>WIN32</Defines>
--- a/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2019.vcxproj
+++ b/Samples/cannyEdgeDetectorNPP/cannyEdgeDetectorNPP_vs2019.vcxproj
@@ -63,7 +63,7 @@
      <OutputFile>$(OutDir)/cannyEdgeDetectorNPP.exe</OutputFile>
    </Link>
    <CudaCompile>
-      <CodeGeneration>compute_30,compute_30;</CodeGeneration>
+      <CodeGeneration>compute_35,compute_35;</CodeGeneration>
      <AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
      <Include>./;../../Common</Include>
      <Defines>WIN32</Defines>