From c9811fca37d3591e9377e9362ba6eb129e671b6b Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Thu, 26 Mar 2020 11:07:43 +0300 Subject: [PATCH] [HIPIFY][doc] Update README.md: LLVM 10.0.0 Release is supported [ROCm/clr commit: cc2991f714b1037586226c68ef901209cb625778] --- projects/clr/hipamd/hipify-clang/README.md | 243 +++++++++++---------- 1 file changed, 122 insertions(+), 121 deletions(-) diff --git a/projects/clr/hipamd/hipify-clang/README.md b/projects/clr/hipamd/hipify-clang/README.md index bfa017def2..8cdeceace0 100644 --- a/projects/clr/hipamd/hipify-clang/README.md +++ b/projects/clr/hipamd/hipify-clang/README.md @@ -42,10 +42,9 @@ After applying all the matchers, the output HIP source is produced. `hipify-clang` requires: -1. [**LLVM+CLANG**](http://releases.llvm.org) of at least version [3.8.0](http://releases.llvm.org/download.html#3.8.0); the latest stable and recommended release: [**9.0.1**](http://releases.llvm.org/download.html#9.0.1), the latest release candidate: [10.0.0-rc4](https://github.com/llvm/llvm-project/releases/tag/llvmorg-10.0.0-rc4). +1. [**LLVM+CLANG**](http://releases.llvm.org) of at least version [3.8.0](http://releases.llvm.org/download.html#3.8.0); the latest stable and recommended release: [**10.0.0**](http://releases.llvm.org/download.html#10.0.0). -2. [**CUDA**](https://developer.nvidia.com/cuda-downloads) of at least version [7.0](https://developer.nvidia.com/cuda-toolkit-70), the latest supported version is [**10.1 Update 2**](https://developer.nvidia.com/cuda-10.1-download-archive-base). -To use the latest CUDA version [10.2](https://developer.nvidia.com/cuda-downloads) please use the latest `LLVM` release candidate: [10.0.0-rc4](https://github.com/llvm/llvm-project/releases/tag/llvmorg-10.0.0-rc4). +2. [**CUDA**](https://developer.nvidia.com/cuda-downloads) of at least version [7.0](https://developer.nvidia.com/cuda-toolkit-70), the latest supported version is [**10.2**](https://developer.nvidia.com/cuda-downloads). | **LLVM release version** | **CUDA latest supported version** | **Windows** | **Linux** | |:----------------------------------------------------------:|:------------------------------------------------------------------------:|:-----------:|:---------:| @@ -66,15 +65,15 @@ To use the latest CUDA version [10.2](https://developer.nvidia.com/cuda-download | [8.0.0](http://releases.llvm.org/download.html#8.0.0) | [10.0](https://developer.nvidia.com/cuda-10.0-download-archive) | -
not working due to
the clang's bug [38811](https://bugs.llvm.org/show_bug.cgi?id=38811)
+
[patch](patches/patch_for_clang_8.0.0_bug_38811.zip)*
| + | | [8.0.1](http://releases.llvm.org/download.html#8.0.1) | [10.0](https://developer.nvidia.com/cuda-10.0-download-archive) | -
not working due to
the clang's bug [38811](https://bugs.llvm.org/show_bug.cgi?id=38811)
+
[patch](patches/patch_for_clang_8.0.1_bug_38811.zip)*
| + | | [9.0.0](http://releases.llvm.org/download.html#9.0.0) | [10.1](https://developer.nvidia.com/cuda-10.1-download-archive-base) | + | + | -| [**9.0.1**](http://releases.llvm.org/download.html#9.0.1) | [**10.1**](https://developer.nvidia.com/cuda-10.1-download-archive-base) | +
**LATEST STABLE RELEASE** | +
**LATEST STABLE RELEASE** | -| [10.0.0-rc4](https://github.com/llvm/llvm-project/releases/tag/llvmorg-10.0.0-rc4) | [10.2](https://developer.nvidia.com/cuda-downloads) | + | + | +| [9.0.1](http://releases.llvm.org/download.html#9.0.1) | [10.1](https://developer.nvidia.com/cuda-10.1-download-archive-base) | + | + | +| [**10.0.0**](http://releases.llvm.org/download.html#10.0.0)| [**10.2**](https://developer.nvidia.com/cuda-downloads) | +
**LATEST STABLE RELEASE** | +
**LATEST STABLE RELEASE** | `*` Download the patch and unpack it into your `LLVM` distributive directory; a few header files will be overwritten; rebuilding of `LLVM` is not needed. In most cases, you can get a suitable version of `LLVM+CLANG` with your package manager. Failing that or having multiple versions of `LLVM`, you can [download a release archive](http://releases.llvm.org/), build or install it, and set -[CMAKE_PREFIX_PATH](https://cmake.org/cmake/help/v3.5/variable/CMAKE_PREFIX_PATH.html) so `cmake` can find it; for instance: `-DCMAKE_PREFIX_PATH=f:\LLVM\9.0.1\dist` +[CMAKE_PREFIX_PATH](https://cmake.org/cmake/help/v3.5/variable/CMAKE_PREFIX_PATH.html) so `cmake` can find it; for instance: `-DCMAKE_PREFIX_PATH=d:\LLVM\10.0.0\dist` ### hipify-clang: usage @@ -83,14 +82,14 @@ To process a file, `hipify-clang` needs access to the same headers that would be For example: ```shell -./hipify-clang square.cu --cuda-path=/usr/local/cuda-10.1 -I /usr/local/cuda-10.1/samples/common/inc +./hipify-clang square.cu --cuda-path=/usr/local/cuda-10.2 -I /usr/local/cuda-10.2/samples/common/inc ``` `hipify-clang` arguments are given first, followed by a separator `'--'`, and then the arguments you'd pass to `clang` if you were compiling the input file. For example: ```bash -./hipify-clang cpp17.cu --cuda-path=/usr/local/cuda-10.1 -- -std=c++17 +./hipify-clang cpp17.cu --cuda-path=/usr/local/cuda-10.2 -- -std=c++17 ``` The [Clang manual for compiling CUDA](https://llvm.org/docs/CompileCudaWithLLVM.html#compiling-cuda-code) may be useful. @@ -158,7 +157,7 @@ Run `Visual Studio 16 2019`, open the generated `LLVM.sln`, build all, build pro **LLVM 10.0.0 or newer:** -1. download [`LLVM project`](https://github.com/llvm/llvm-project/archive/llvmorg-10.0.0-rc4.tar.gz) sources; +1. download [`LLVM project`](https://github.com/llvm/llvm-project/releases/download/llvmorg-10.0.0/llvm-project-10.0.0.tar.xz) sources; 2. build [`LLVM project`](http://llvm.org/docs/CMake.html): **Linux**: @@ -193,19 +192,19 @@ Run `Visual Studio 16 2019`, open the generated `LLVM.sln`, build all, build pro * Having multiple CUDA installations to choose a particular version the `DCUDA_TOOLKIT_ROOT_DIR` option should be specified: - - ***Linux***: `-DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-10.1` + - ***Linux***: `-DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-10.2` - - ***Windows***: `-DCUDA_TOOLKIT_ROOT_DIR="c:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.1"` + - ***Windows***: `-DCUDA_TOOLKIT_ROOT_DIR="c:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.2"` - `-DCUDA_SDK_ROOT_DIR="c:/ProgramData/NVIDIA Corporation/CUDA Samples/v10.1"` + `-DCUDA_SDK_ROOT_DIR="c:/ProgramData/NVIDIA Corporation/CUDA Samples/v10.2"` 4. Ensure [`cuDNN`](https://developer.nvidia.com/rdp/cudnn-archive) of the version corresponding to CUDA's version is installed. * Path to cuDNN should be specified by the `CUDA_DNN_ROOT_DIR` option: - - ***Linux***: `-DCUDA_DNN_ROOT_DIR=/srv/CUDNN/cudnn-10.1-v7.6.5.32` + - ***Linux***: `-DCUDA_DNN_ROOT_DIR=/srv/CUDNN/cudnn-10.2-v7.6.5.32` - - ***Windows***: `-DCUDA_DNN_ROOT_DIR=f:/CUDNN/cudnn-10.1-windows10-x64-v7.6.5.32` + - ***Windows***: `-DCUDA_DNN_ROOT_DIR=d:/CUDNN/cudnn-10.2-windows10-x64-v7.6.5.32` 5. Ensure [`CUB`](https://github.com/NVlabs/cub) of the version corresponding to CUDA's version is installed. @@ -213,7 +212,7 @@ Run `Visual Studio 16 2019`, open the generated `LLVM.sln`, build all, build pro - ***Linux***: `-DCUDA_CUB_ROOT_DIR=/srv/git/CUB` - - ***Windows***: `-DCUDA_CUB_ROOT_DIR=f:/GIT/cub` + - ***Windows***: `-DCUDA_CUB_ROOT_DIR=d:/GIT/cub` 5. Ensure [`python`](https://www.python.org/downloads) of minimum required version 2.7 is installed. @@ -221,21 +220,21 @@ Run `Visual Studio 16 2019`, open the generated `LLVM.sln`, build all, build pro * Install `lit` into `python`: - - ***Linux***: `python /srv/git/LLVM/9.0.1/llvm/utils/lit/setup.py install` + - ***Linux***: `python /srv/git/LLVM/10.0.0/llvm/utils/lit/setup.py install` - - ***Windows***: `python f:/LLVM/9.0.1/llvm/utils/lit/setup.py install` + - ***Windows***: `python d:/LLVM/10.0.0/llvm/utils/lit/setup.py install` * Starting with LLVM 6.0.1 path to `llvm-lit` python script should be specified by the `LLVM_EXTERNAL_LIT` option: - - ***Linux***: `-DLLVM_EXTERNAL_LIT=/srv/git/LLVM/9.0.1/build/bin/llvm-lit` + - ***Linux***: `-DLLVM_EXTERNAL_LIT=/srv/git/LLVM/10.0.0/build/bin/llvm-lit` - - ***Windows***: `-DLLVM_EXTERNAL_LIT=f:/LLVM/9.0.1/build/Release/bin/llvm-lit.py` + - ***Windows***: `-DLLVM_EXTERNAL_LIT=d:/LLVM/10.0.0/build/Release/bin/llvm-lit.py` * `FileCheck`: - - ***Linux***: copy from `/srv/git/LLVM/9.0.1/build/bin/` to `CMAKE_INSTALL_PREFIX/dist/bin` + - ***Linux***: copy from `/srv/git/LLVM/10.0.0/build/bin/` to `CMAKE_INSTALL_PREFIX/dist/bin` - - ***Windows***: copy from `f:/LLVM/9.0.1/build/Release/bin` to `CMAKE_INSTALL_PREFIX/dist/bin` + - ***Windows***: copy from `d:/LLVM/10.0.0/build/Release/bin` to `CMAKE_INSTALL_PREFIX/dist/bin` - Or specify the path to `FileCheck` in `CMAKE_INSTALL_PREFIX` option @@ -249,7 +248,7 @@ On Linux the following configurations are tested: Ubuntu 14: LLVM 5.0.0 - 6.0.1, CUDA 7.0 - 9.0, cudnn-5.0.5 - cudnn-7.6.5.32 -Ubuntu 16-18: LLVM 8.0.0 - 10.0.0-rc4, CUDA 8.0 - 10.2, cudnn-5.1.10 - cudnn-7.6.5.32 +Ubuntu 16-18: LLVM 8.0.0 - 10.0.0, CUDA 8.0 - 10.2, cudnn-5.1.10 - cudnn-7.6.5.32 Minimum build system requirements for the above configurations: @@ -262,11 +261,11 @@ cmake -DHIPIFY_CLANG_TESTS=1 \ -DCMAKE_BUILD_TYPE=Release \ -DCMAKE_INSTALL_PREFIX=../dist \ - -DCMAKE_PREFIX_PATH=/srv/git/LLVM/9.0.1/dist \ - -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-10.1 \ - -DCUDA_DNN_ROOT_DIR=/srv/CUDNN/cudnn-10.1-v7.6.5.32 \ + -DCMAKE_PREFIX_PATH=/srv/git/LLVM/10.0.0/dist \ + -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-10.2 \ + -DCUDA_DNN_ROOT_DIR=/srv/CUDNN/cudnn-10.2-v7.6.5.32 \ -DCUDA_CUB_ROOT_DIR=/srv/git/CUB \ - -DLLVM_EXTERNAL_LIT=/srv/git/LLVM/9.0.1/build/bin/llvm-lit \ + -DLLVM_EXTERNAL_LIT=/srv/git/LLVM/10.0.0/build/bin/llvm-lit \ .. ``` *A corresponding successful output:* @@ -285,14 +284,14 @@ cmake -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done --- Found LLVM 9.0.1: --- - CMake module path: /srv/git/LLVM/9.0.1/dist/lib/cmake/llvm --- - Include path : /srv/git/LLVM/9.0.1/dist/include --- - Binary path : /srv/git/LLVM/9.0.1/dist/bin +-- Found LLVM 10.0.0: +-- - CMake module path: /srv/git/LLVM/10.0.0/dist/lib/cmake/llvm +-- - Include path : /srv/git/LLVM/10.0.0/dist/include +-- - Binary path : /srv/git/LLVM/10.0.0/dist/bin -- Linker detection: GNU ld -- Found PythonInterp: /usr/bin/python2.7 (found suitable version "2.7.12", minimum required is "2.7") -- Found lit: /usr/local/bin/lit --- Found FileCheck: /srv/git/LLVM/9.0.1/dist/bin/FileCheck +-- Found FileCheck: /srv/git/LLVM/10.0.0/dist/bin/FileCheck -- Looking for pthread.h -- Looking for pthread.h - found -- Looking for pthread_create @@ -302,7 +301,7 @@ cmake -- Looking for pthread_create in pthread -- Looking for pthread_create in pthread - found -- Found Threads: TRUE --- Found CUDA: /usr/local/cuda-10.1 (found version "10.1") +-- Found CUDA: /usr/local/cuda-10.2 (found version "10.2") -- Configuring done -- Generating done -- Build files have been written to: /srv/git/HIP/hipify-clang/build @@ -314,83 +313,85 @@ make test-hipify ```shell Running HIPify regression tests ======================================== -CUDA 10.1 - will be used for testing -LLVM 9.0.1 - will be used for testing +CUDA 10.2 - will be used for testing +LLVM 10.0.0 - will be used for testing x86_64 - Platform architecture Linux 5.2.0 - Platform OS 64 - hipify-clang binary bitness 64 - python 2.7.12 binary bitness ======================================== --- Testing: 67 tests, 12 threads -- -PASS: hipify :: unit_tests/casts/reinterpret_cast.cu (1 of 67) -PASS: hipify :: unit_tests/device/math_functions.cu (2 of 67) -PASS: hipify :: unit_tests/device/atomics.cu (3 of 67) -PASS: hipify :: unit_tests/device/device_symbols.cu (4 of 67) -PASS: hipify :: unit_tests/headers/headers_test_01.cu (5 of 67) -PASS: hipify :: unit_tests/headers/headers_test_02.cu (6 of 67) -PASS: hipify :: unit_tests/headers/headers_test_03.cu (7 of 67) -PASS: hipify :: unit_tests/headers/headers_test_05.cu (8 of 67) -PASS: hipify :: unit_tests/headers/headers_test_04.cu (9 of 67) -PASS: hipify :: unit_tests/headers/headers_test_06.cu (10 of 67) -PASS: hipify :: unit_tests/headers/headers_test_07.cu (11 of 67) -PASS: hipify :: unit_tests/headers/headers_test_10.cu (12 of 67) -PASS: hipify :: unit_tests/headers/headers_test_11.cu (13 of 67) -PASS: hipify :: unit_tests/headers/headers_test_08.cu (14 of 67) -PASS: hipify :: unit_tests/kernel_launch/kernel_launch_01.cu (15 of 67) -PASS: hipify :: unit_tests/headers/headers_test_09.cu (16 of 67) -PASS: hipify :: unit_tests/libraries/CAFFE2/caffe2_02.cu (17 of 67) -PASS: hipify :: unit_tests/libraries/CAFFE2/caffe2_01.cu (18 of 67) -PASS: hipify :: unit_tests/libraries/cuBLAS/cublas_0_based_indexing.cu (19 of 67) -PASS: hipify :: unit_tests/libraries/cuBLAS/cublas_1_based_indexing.cu (20 of 67) -PASS: hipify :: unit_tests/libraries/CUB/cub_03.cu (21 of 67) -PASS: hipify :: unit_tests/libraries/CUB/cub_01.cu (22 of 67) -PASS: hipify :: unit_tests/libraries/CUB/cub_02.cu (23 of 67) -PASS: hipify :: unit_tests/libraries/cuBLAS/rocBLAS/cublas_0_based_indexing_rocblas.cu (24 of 67) -PASS: hipify :: unit_tests/libraries/cuBLAS/cublas_sgemm_matrix_multiplication.cu (25 of 67) -PASS: hipify :: unit_tests/libraries/cuBLAS/rocBLAS/cublas_1_based_indexing_rocblas.cu (26 of 67) -PASS: hipify :: unit_tests/libraries/cuBLAS/rocBLAS/cublas_sgemm_matrix_multiplication_rocblas.cu (27 of 67) -PASS: hipify :: unit_tests/libraries/cuComplex/cuComplex_Julia.cu (28 of 67) -PASS: hipify :: unit_tests/libraries/cuFFT/simple_cufft.cu (29 of 67) -PASS: hipify :: unit_tests/libraries/cuDNN/cudnn_softmax.cu (30 of 67) -PASS: hipify :: unit_tests/libraries/cuDNN/cudnn_convolution_forward.cu (31 of 67) -PASS: hipify :: unit_tests/libraries/cuRAND/poisson_api_example.cu (32 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_01.cu (33 of 67) -PASS: hipify :: unit_tests/libraries/cuRAND/benchmark_curand_generate.cpp (34 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_02.cu (35 of 67) -PASS: hipify :: unit_tests/libraries/cuRAND/benchmark_curand_kernel.cpp (36 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_03.cu (37 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_04.cu (38 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_05.cu (39 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_07.cu (40 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_06.cu (41 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_08.cu (42 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_09.cu (43 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_11.cu (44 of 67) -PASS: hipify :: unit_tests/namespace/ns_kernel_launch.cu (45 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_10.cu (46 of 67) -PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_12.cu (47 of 67) -PASS: hipify :: unit_tests/pp/pp_if_else_conditionals.cu (48 of 67) -PASS: hipify :: unit_tests/pp/pp_if_else_conditionals_01.cu (49 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/11_texture_driver/tex2dKernel.cpp (50 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/0_MatrixTranspose/MatrixTranspose.cpp (51 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/11_texture_driver/texture2dDrv.cpp (52 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/13_occupancy/occupancy.cpp (53 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/1_hipEvent/hipEvent.cpp (54 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/2_Profiler/Profiler.cpp (55 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/7_streams/stream.cpp (56 of 67) -PASS: hipify :: unit_tests/samples/2_Cookbook/8_peer2peer/peer2peer.cpp (57 of 67) -PASS: hipify :: unit_tests/samples/MallocManaged.cpp (58 of 67) -PASS: hipify :: unit_tests/samples/allocators.cu (59 of 67) -PASS: hipify :: unit_tests/samples/coalescing.cu (60 of 67) -PASS: hipify :: unit_tests/samples/dynamic_shared_memory.cu (61 of 67) -PASS: hipify :: unit_tests/samples/axpy.cu (62 of 67) -PASS: hipify :: unit_tests/samples/intro.cu (63 of 67) -PASS: hipify :: unit_tests/samples/cudaRegister.cu (64 of 67) -PASS: hipify :: unit_tests/samples/square.cu (65 of 67) -PASS: hipify :: unit_tests/samples/static_shared_memory.cu (66 of 67) -PASS: hipify :: unit_tests/samples/vec_add.cu (67 of 67) -Testing Time: 3.07s - Expected Passes : 67 +-- Testing: 69 tests, 12 threads -- +PASS: hipify :: unit_tests/casts/reinterpret_cast.cu (1 of 69) +PASS: hipify :: unit_tests/device/math_functions.cu (2 of 69) +PASS: hipify :: unit_tests/device/atomics.cu (3 of 69) +PASS: hipify :: unit_tests/headers/headers_test_01.cu (4 of 69) +PASS: hipify :: unit_tests/device/device_symbols.cu (5 of 69) +PASS: hipify :: unit_tests/headers/headers_test_02.cu (6 of 69) +PASS: hipify :: unit_tests/headers/headers_test_03.cu (7 of 69) +PASS: hipify :: unit_tests/headers/headers_test_05.cu (8 of 69) +PASS: hipify :: unit_tests/headers/headers_test_04.cu (9 of 69) +PASS: hipify :: unit_tests/headers/headers_test_07.cu (10 of 69) +PASS: hipify :: unit_tests/headers/headers_test_06.cu (11 of 69) +PASS: hipify :: unit_tests/headers/headers_test_11.cu (12 of 69) +PASS: hipify :: unit_tests/headers/headers_test_10.cu (13 of 69) +PASS: hipify :: unit_tests/headers/headers_test_08.cu (14 of 69) +PASS: hipify :: unit_tests/kernel_launch/kernel_launch_01.cu (15 of 69) +PASS: hipify :: unit_tests/libraries/CAFFE2/caffe2_02.cu (16 of 69) +PASS: hipify :: unit_tests/headers/headers_test_09.cu (17 of 69) +PASS: hipify :: unit_tests/libraries/CAFFE2/caffe2_01.cu (18 of 69) +PASS: hipify :: unit_tests/libraries/cuBLAS/cublas_0_based_indexing.cu (19 of 69) +PASS: hipify :: unit_tests/libraries/cuBLAS/cublas_1_based_indexing.cu (20 of 69) +PASS: hipify :: unit_tests/libraries/CUB/cub_03.cu (21 of 69) +PASS: hipify :: unit_tests/libraries/CUB/cub_01.cu (22 of 69) +PASS: hipify :: unit_tests/libraries/CUB/cub_02.cu (23 of 69) +PASS: hipify :: unit_tests/libraries/cuBLAS/cublas_sgemm_matrix_multiplication.cu (24 of 69) +PASS: hipify :: unit_tests/libraries/cuBLAS/rocBLAS/cublas_0_based_indexing_rocblas.cu (25 of 69) +PASS: hipify :: unit_tests/libraries/cuBLAS/rocBLAS/cublas_1_based_indexing_rocblas.cu (26 of 69) +PASS: hipify :: unit_tests/libraries/cuBLAS/rocBLAS/cublas_sgemm_matrix_multiplication_rocblas.cu (27 of 69) +PASS: hipify :: unit_tests/libraries/cuComplex/cuComplex_Julia.cu (28 of 69) +PASS: hipify :: unit_tests/libraries/cuDNN/cudnn_softmax.cu (29 of 69) +PASS: hipify :: unit_tests/libraries/cuFFT/simple_cufft.cu (30 of 69) +PASS: hipify :: unit_tests/libraries/cuDNN/cudnn_convolution_forward.cu (31 of 69) +PASS: hipify :: unit_tests/libraries/cuRAND/poisson_api_example.cu (32 of 69) +PASS: hipify :: unit_tests/libraries/cuRAND/benchmark_curand_generate.cpp (33 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_01.cu (34 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_02.cu (35 of 69) +PASS: hipify :: unit_tests/libraries/cuRAND/benchmark_curand_kernel.cpp (36 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_03.cu (37 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_04.cu (38 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_05.cu (39 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_06.cu (40 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_07.cu (41 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_08.cu (42 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_09.cu (43 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_10.cu (44 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_11.cu (45 of 69) +PASS: hipify :: unit_tests/namespace/ns_kernel_launch.cu (46 of 69) +PASS: hipify :: unit_tests/libraries/cuSPARSE/cuSPARSE_12.cu (47 of 69) +PASS: hipify :: unit_tests/pp/pp_if_else_conditionals.cu (48 of 69) +PASS: hipify :: unit_tests/pp/pp_if_else_conditionals_01.cu (49 of 69) +PASS: hipify :: unit_tests/pp/pp_if_else_conditionals_01_LLVM_10.cu (50 of 69) +PASS: hipify :: unit_tests/pp/pp_if_else_conditionals_LLVM_10.cu (51 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/11_texture_driver/tex2dKernel.cpp (52 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/0_MatrixTranspose/MatrixTranspose.cpp (53 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/11_texture_driver/texture2dDrv.cpp (54 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/1_hipEvent/hipEvent.cpp (55 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/13_occupancy/occupancy.cpp (56 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/2_Profiler/Profiler.cpp (57 of 69) +PASS: hipify :: unit_tests/samples/MallocManaged.cpp (58 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/7_streams/stream.cpp (59 of 69) +PASS: hipify :: unit_tests/samples/2_Cookbook/8_peer2peer/peer2peer.cpp (60 of 69) +PASS: hipify :: unit_tests/samples/allocators.cu (61 of 69) +PASS: hipify :: unit_tests/samples/coalescing.cu (62 of 69) +PASS: hipify :: unit_tests/samples/axpy.cu (63 of 69) +PASS: hipify :: unit_tests/samples/dynamic_shared_memory.cu (64 of 69) +PASS: hipify :: unit_tests/samples/cudaRegister.cu (65 of 69) +PASS: hipify :: unit_tests/samples/intro.cu (66 of 69) +PASS: hipify :: unit_tests/samples/square.cu (67 of 69) +PASS: hipify :: unit_tests/samples/static_shared_memory.cu (68 of 69) +PASS: hipify :: unit_tests/samples/vec_add.cu (69 of 69) +Testing Time: 3.23s + Expected Passes : 69 [100%] Built target test-hipify ``` ### hipify-clang: Windows @@ -404,8 +405,8 @@ Testing Time: 3.07s | 7.0.0 - 7.1.0 | 9.2 | 7.6.5.32 | 2017.15.9.11 | 3.13.3 | 3.7.3 | | 8.0.0 - 8.0.1 | 10.0 | 7.6.5.32 | 2017.15.9.15 | 3.14.2 | 3.7.4 | | 9.0.0 - 9.0.1 | 10.1 | 7.6.5.32 | 2017.15.9.20, 2019.16.4.5 | 3.16.4 | 3.8.0 | -| 10.0.0-rc1-rc4 | 10.2 | 7.6.5.32 | 2017.15.9.21, 2019.16.5.0 | 3.16.5 | 3.8.2 | -| 11.0.0git | 10.2 | 7.6.5.32 | 2017.15.9.21, 2019.16.5.0 | 3.16.5 | 3.8.2 | +| 10.0.0 | 10.2 | 7.6.5.32 | 2017.15.9.21, 2019.16.5.1 | 3.17.0 | 3.8.2 | +| 11.0.0git | 10.2 | 7.6.5.32 | 2017.15.9.21, 2019.16.5.1 | 3.17.0 | 3.8.2 | *Building with testing support on `Windows 10` by `Visual Studio 16 2019`:* @@ -416,28 +417,28 @@ cmake -DHIPIFY_CLANG_TESTS=1 \ -DCMAKE_BUILD_TYPE=Release \ -DCMAKE_INSTALL_PREFIX=../dist \ - -DCMAKE_PREFIX_PATH=f:/LLVM/9.0.1/dist \ - -DCUDA_TOOLKIT_ROOT_DIR="c:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.1" \ - -DCUDA_SDK_ROOT_DIR="c:/ProgramData/NVIDIA Corporation/CUDA Samples/v10.1" \ - -DCUDA_DNN_ROOT_DIR=f:/CUDNN/cudnn-10.1-windows10-x64-v7.6.5.32 \ - -DCUDA_CUB_ROOT_DIR=f:/GIT/cub \ - -DLLVM_EXTERNAL_LIT=f:/LLVM/9.0.1/build/Release/bin/llvm-lit.py \ + -DCMAKE_PREFIX_PATH=d:/LLVM/10.0.0/dist \ + -DCUDA_TOOLKIT_ROOT_DIR="c:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.2" \ + -DCUDA_SDK_ROOT_DIR="c:/ProgramData/NVIDIA Corporation/CUDA Samples/v10.2" \ + -DCUDA_DNN_ROOT_DIR=d:/CUDNN/cudnn-10.2-windows10-x64-v7.6.5.32 \ + -DCUDA_CUB_ROOT_DIR=d:/GIT/cub \ + -DLLVM_EXTERNAL_LIT=d:/LLVM/10.0.0/build/Release/bin/llvm-lit.py \ -Thost=x64 .. ``` *A corresponding successful output:* ```shell --- Found LLVM 9.0.1: --- - CMake module path: F:/LLVM/9.0.1/dist/lib/cmake/llvm --- - Include path : F:/LLVM/9.0.1/dist/include --- - Binary path : F:/LLVM/9.0.1/dist/bin --- Found PythonInterp: C:/Program Files/Python38/python.exe (found suitable version "3.8.2", minimum required is "3.6") --- Found lit: C:/Program Files/Python38/Scripts/lit.exe --- Found FileCheck: F:/LLVM/9.0.1/dist/bin/FileCheck.exe --- Found CUDA: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.1 (found version "10.1") +-- Found LLVM 10.0.0: +-- - CMake module path: d:/LLVM/10.0.0/dist/lib/cmake/llvm +-- - Include path : d:/LLVM/10.0.0/dist/include +-- - Binary path : d:/LLVM/10.0.0/dist/bin +-- Found PythonInterp: c:/Program Files/Python38/python.exe (found suitable version "3.8.2", minimum required is "3.6") +-- Found lit: c:/Program Files/Python38/Scripts/lit.exe +-- Found FileCheck: d:/LLVM/10.0.0/dist/bin/FileCheck.exe +-- Found CUDA: c:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.2 (found version "10.2") -- Configuring done -- Generating done --- Build files have been written to: f:/HIP/hipify-clang/build +-- Build files have been written to: d:/HIP/hipify-clang/build ``` Run `Visual Studio 16 2019`, open the generated `hipify-clang.sln`, build project `test-hipify`.