mirror of
https://github.com/pytorch/pytorch.git
synced 2025-10-20 21:14:14 +08:00
[BE] Remove CUDA 11 artifacts. Fix Check Binary workflow (#155555)
Please see: https://github.com/pytorch/pytorch/issues/147383 1. Remove CUDA 11 build and test artifacts. One place CUDA 12.4 2. Fix Check Binary Workflow to use Stable Cuda version variable rather then hardcoded one Pull Request resolved: https://github.com/pytorch/pytorch/pull/155555 Approved by: https://github.com/malfet, https://github.com/Skylion007
This commit is contained in:
atalman
committed by
PyTorch MergeBot
PyTorch MergeBot
parent
40fefe2871
commit
7a03b0d2ca
@ -52,10 +52,6 @@ ENV CUDA_VERSION=${CUDA_VERSION}
|
||||
# Make things in our path by default
|
||||
ENV PATH=/usr/local/cuda-${CUDA_VERSION}/bin:$PATH
|
||||
|
||||
FROM cuda as cuda11.8
|
||||
RUN bash ./install_cuda.sh 11.8
|
||||
ENV DESIRED_CUDA=11.8
|
||||
|
||||
FROM cuda as cuda12.6
|
||||
RUN bash ./install_cuda.sh 12.6
|
||||
ENV DESIRED_CUDA=12.6
|
||||
|
||||
@ -30,16 +30,6 @@ install_ubuntu() {
|
||||
maybe_libomp_dev=""
|
||||
fi
|
||||
|
||||
# HACK: UCC testing relies on libnccl library from NVIDIA repo, and version 2.16 crashes
|
||||
# See https://github.com/pytorch/pytorch/pull/105260#issuecomment-1673399729
|
||||
# TODO: Eliminate this hack, we should not relay on apt-get installation
|
||||
# See https://github.com/pytorch/pytorch/issues/144768
|
||||
if [[ "$UBUNTU_VERSION" == "20.04"* && "$CUDA_VERSION" == "11.8"* ]]; then
|
||||
maybe_libnccl_dev="libnccl2=2.15.5-1+cuda11.8 libnccl-dev=2.15.5-1+cuda11.8 --allow-downgrades --allow-change-held-packages"
|
||||
else
|
||||
maybe_libnccl_dev=""
|
||||
fi
|
||||
|
||||
# Install common dependencies
|
||||
apt-get update
|
||||
# TODO: Some of these may not be necessary
|
||||
@ -68,7 +58,6 @@ install_ubuntu() {
|
||||
libasound2-dev \
|
||||
libsndfile-dev \
|
||||
${maybe_libomp_dev} \
|
||||
${maybe_libnccl_dev} \
|
||||
software-properties-common \
|
||||
wget \
|
||||
sudo \
|
||||
|
||||
@ -40,20 +40,6 @@ function install_cudnn {
|
||||
rm -rf tmp_cudnn
|
||||
}
|
||||
|
||||
function install_118 {
|
||||
CUDNN_VERSION=9.1.0.70
|
||||
echo "Installing CUDA 11.8 and cuDNN ${CUDNN_VERSION} and NCCL and cuSparseLt-0.4.0"
|
||||
install_cuda 11.8.0 cuda_11.8.0_520.61.05_linux
|
||||
|
||||
install_cudnn 11 $CUDNN_VERSION
|
||||
|
||||
CUDA_VERSION=11.8 bash install_nccl.sh
|
||||
|
||||
CUDA_VERSION=11.8 bash install_cusparselt.sh
|
||||
|
||||
ldconfig
|
||||
}
|
||||
|
||||
function install_126 {
|
||||
CUDNN_VERSION=9.5.1.17
|
||||
echo "Installing CUDA 12.6.3 and cuDNN ${CUDNN_VERSION} and NCCL and cuSparseLt-0.7.1"
|
||||
@ -84,37 +70,6 @@ function install_129 {
|
||||
ldconfig
|
||||
}
|
||||
|
||||
function prune_118 {
|
||||
echo "Pruning CUDA 11.8 and cuDNN"
|
||||
#####################################################################################
|
||||
# CUDA 11.8 prune static libs
|
||||
#####################################################################################
|
||||
export NVPRUNE="/usr/local/cuda-11.8/bin/nvprune"
|
||||
export CUDA_LIB_DIR="/usr/local/cuda-11.8/lib64"
|
||||
|
||||
export GENCODE="-gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_90,code=sm_90"
|
||||
export GENCODE_CUDNN="-gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_90,code=sm_90"
|
||||
|
||||
if [[ -n "$OVERRIDE_GENCODE" ]]; then
|
||||
export GENCODE=$OVERRIDE_GENCODE
|
||||
fi
|
||||
|
||||
# all CUDA libs except CuDNN and CuBLAS (cudnn and cublas need arch 3.7 included)
|
||||
ls $CUDA_LIB_DIR/ | grep "\.a" | grep -v "culibos" | grep -v "cudart" | grep -v "cudnn" | grep -v "cublas" | grep -v "metis" \
|
||||
| xargs -I {} bash -c \
|
||||
"echo {} && $NVPRUNE $GENCODE $CUDA_LIB_DIR/{} -o $CUDA_LIB_DIR/{}"
|
||||
|
||||
# prune CuDNN and CuBLAS
|
||||
$NVPRUNE $GENCODE_CUDNN $CUDA_LIB_DIR/libcublas_static.a -o $CUDA_LIB_DIR/libcublas_static.a
|
||||
$NVPRUNE $GENCODE_CUDNN $CUDA_LIB_DIR/libcublasLt_static.a -o $CUDA_LIB_DIR/libcublasLt_static.a
|
||||
|
||||
#####################################################################################
|
||||
# CUDA 11.8 prune visual tools
|
||||
#####################################################################################
|
||||
export CUDA_BASE="/usr/local/cuda-11.8/"
|
||||
rm -rf $CUDA_BASE/libnvvp $CUDA_BASE/nsightee_plugins $CUDA_BASE/nsight-compute-2022.3.0 $CUDA_BASE/nsight-systems-2022.4.2/
|
||||
}
|
||||
|
||||
function prune_126 {
|
||||
echo "Pruning CUDA 12.6"
|
||||
#####################################################################################
|
||||
@ -169,8 +124,6 @@ function install_128 {
|
||||
while test $# -gt 0
|
||||
do
|
||||
case "$1" in
|
||||
11.8) install_118; prune_118
|
||||
;;
|
||||
12.6|12.6.*) install_126; prune_126
|
||||
;;
|
||||
12.8|12.8.*) install_128;
|
||||
|
||||
@ -13,9 +13,6 @@ if [[ ${CUDA_VERSION:0:4} =~ ^12\.[5-9]$ ]]; then
|
||||
fi
|
||||
CUSPARSELT_NAME="libcusparse_lt-linux-${arch_path}-0.7.1.0-archive"
|
||||
curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cusparselt/redist/libcusparse_lt/linux-${arch_path}/${CUSPARSELT_NAME}.tar.xz
|
||||
elif [[ ${CUDA_VERSION:0:4} == "11.8" ]]; then
|
||||
CUSPARSELT_NAME="libcusparse_lt-linux-x86_64-0.4.0.7-archive"
|
||||
curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cusparselt/redist/libcusparse_lt/linux-x86_64/${CUSPARSELT_NAME}.tar.xz
|
||||
else
|
||||
echo "Not sure which libcusparselt version to install for this ${CUDA_VERSION}"
|
||||
fi
|
||||
|
||||
@ -54,16 +54,6 @@ COPY ./ci_commit_pins/nccl-cu* /ci_commit_pins/
|
||||
COPY ./common/install_cusparselt.sh install_cusparselt.sh
|
||||
ENV CUDA_HOME /usr/local/cuda
|
||||
|
||||
FROM cuda as cuda11.8
|
||||
RUN bash ./install_cuda.sh 11.8
|
||||
RUN bash ./install_magma.sh 11.8
|
||||
RUN ln -sf /usr/local/cuda-11.8 /usr/local/cuda
|
||||
|
||||
FROM cuda as cuda12.4
|
||||
RUN bash ./install_cuda.sh 12.4
|
||||
RUN bash ./install_magma.sh 12.4
|
||||
RUN ln -sf /usr/local/cuda-12.4 /usr/local/cuda
|
||||
|
||||
FROM cuda as cuda12.6
|
||||
RUN bash ./install_cuda.sh 12.6
|
||||
RUN bash ./install_magma.sh 12.6
|
||||
|
||||
@ -1,7 +1,7 @@
|
||||
SHELL=/usr/bin/env bash
|
||||
|
||||
DOCKER_CMD ?= docker
|
||||
DESIRED_CUDA ?= 11.8
|
||||
DESIRED_CUDA ?= 12.8
|
||||
DESIRED_CUDA_SHORT = $(subst .,,$(DESIRED_CUDA))
|
||||
PACKAGE_NAME = magma-cuda
|
||||
CUDA_ARCH_LIST ?= -gencode arch=compute_50,code=sm_50 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_90,code=sm_90
|
||||
@ -19,7 +19,6 @@ DOCKER_RUN = set -eou pipefail; ${DOCKER_CMD} run --rm -i \
|
||||
all: magma-cuda129
|
||||
all: magma-cuda128
|
||||
all: magma-cuda126
|
||||
all: magma-cuda118
|
||||
|
||||
.PHONY:
|
||||
clean:
|
||||
@ -42,9 +41,3 @@ magma-cuda128:
|
||||
magma-cuda126: DESIRED_CUDA := 12.6
|
||||
magma-cuda126:
|
||||
$(DOCKER_RUN)
|
||||
|
||||
.PHONY: magma-cuda118
|
||||
magma-cuda118: DESIRED_CUDA := 11.8
|
||||
magma-cuda118: CUDA_ARCH_LIST += -gencode arch=compute_37,code=sm_37
|
||||
magma-cuda118:
|
||||
$(DOCKER_RUN)
|
||||
|
||||
@ -62,10 +62,6 @@ case ${CUDA_VERSION} in
|
||||
TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST};9.0"
|
||||
EXTRA_CAFFE2_CMAKE_FLAGS+=("-DATEN_NO_TEST=ON")
|
||||
;;
|
||||
11.8)
|
||||
TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST};3.7;9.0"
|
||||
EXTRA_CAFFE2_CMAKE_FLAGS+=("-DATEN_NO_TEST=ON")
|
||||
;;
|
||||
*)
|
||||
echo "unknown cuda version $CUDA_VERSION"
|
||||
exit 1
|
||||
@ -181,85 +177,6 @@ if [[ $CUDA_VERSION == 12* ]]; then
|
||||
export USE_CUDA_STATIC_LINK=0
|
||||
export USE_CUPTI_SO=1
|
||||
fi
|
||||
elif [[ $CUDA_VERSION == "11.8" ]]; then
|
||||
export USE_STATIC_CUDNN=0
|
||||
# Turn USE_CUFILE off for CUDA 11.8 since nvidia-cufile-cu11 and 1.9.0.20 are
|
||||
# not available in PYPI
|
||||
export USE_CUFILE=0
|
||||
# Try parallelizing nvcc as well
|
||||
export TORCH_NVCC_FLAGS="-Xfatbin -compress-all --threads 2"
|
||||
# Bundle ptxas into the wheel, see https://github.com/pytorch/pytorch/pull/119750
|
||||
export BUILD_BUNDLE_PTXAS=1
|
||||
|
||||
# CUDA 11.8 have to ship the libcusparseLt.so.0 with the binary
|
||||
# since nvidia-cusparselt-cu11 is not available in PYPI
|
||||
if [[ $USE_CUSPARSELT == "1" ]]; then
|
||||
DEPS_SONAME+=(
|
||||
"libcusparseLt.so.0"
|
||||
)
|
||||
DEPS_LIST+=(
|
||||
"/usr/local/cuda/lib64/libcusparseLt.so.0"
|
||||
)
|
||||
fi
|
||||
|
||||
if [[ -z "$PYTORCH_EXTRA_INSTALL_REQUIREMENTS" ]]; then
|
||||
echo "Bundling with cudnn and cublas."
|
||||
DEPS_LIST+=(
|
||||
"/usr/local/cuda/lib64/libcudnn_adv.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn_cnn.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn_graph.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn_ops.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn_engines_runtime_compiled.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn_engines_precompiled.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn_heuristic.so.9"
|
||||
"/usr/local/cuda/lib64/libcudnn.so.9"
|
||||
"/usr/local/cuda/lib64/libcublas.so.11"
|
||||
"/usr/local/cuda/lib64/libcublasLt.so.11"
|
||||
"/usr/local/cuda/lib64/libcudart.so.11.0"
|
||||
"/usr/local/cuda/lib64/libnvToolsExt.so.1"
|
||||
"/usr/local/cuda/lib64/libnvrtc.so.11.2" # this is not a mistake, it links to more specific cuda version
|
||||
"/usr/local/cuda/lib64/libnvrtc-builtins.so.11.8"
|
||||
)
|
||||
DEPS_SONAME+=(
|
||||
"libcudnn_adv.so.9"
|
||||
"libcudnn_cnn.so.9"
|
||||
"libcudnn_graph.so.9"
|
||||
"libcudnn_ops.so.9"
|
||||
"libcudnn_engines_runtime_compiled.so.9"
|
||||
"libcudnn_engines_precompiled.so.9"
|
||||
"libcudnn_heuristic.so.9"
|
||||
"libcudnn.so.9"
|
||||
"libcublas.so.11"
|
||||
"libcublasLt.so.11"
|
||||
"libcudart.so.11.0"
|
||||
"libnvToolsExt.so.1"
|
||||
"libnvrtc.so.11.2"
|
||||
"libnvrtc-builtins.so.11.8"
|
||||
)
|
||||
else
|
||||
echo "Using nvidia libs from pypi."
|
||||
CUDA_RPATHS=(
|
||||
'$ORIGIN/../../nvidia/cublas/lib'
|
||||
'$ORIGIN/../../nvidia/cuda_cupti/lib'
|
||||
'$ORIGIN/../../nvidia/cuda_nvrtc/lib'
|
||||
'$ORIGIN/../../nvidia/cuda_runtime/lib'
|
||||
'$ORIGIN/../../nvidia/cudnn/lib'
|
||||
'$ORIGIN/../../nvidia/cufft/lib'
|
||||
'$ORIGIN/../../nvidia/curand/lib'
|
||||
'$ORIGIN/../../nvidia/cusolver/lib'
|
||||
'$ORIGIN/../../nvidia/cusparse/lib'
|
||||
'$ORIGIN/../../nvidia/nccl/lib'
|
||||
'$ORIGIN/../../nvidia/nvtx/lib'
|
||||
)
|
||||
CUDA_RPATHS=$(IFS=: ; echo "${CUDA_RPATHS[*]}")
|
||||
export C_SO_RPATH=$CUDA_RPATHS':$ORIGIN:$ORIGIN/lib'
|
||||
export LIB_SO_RPATH=$CUDA_RPATHS':$ORIGIN'
|
||||
export FORCE_RPATH="--force-rpath"
|
||||
export USE_STATIC_NCCL=0
|
||||
export ATEN_STATIC_CUDA=0
|
||||
export USE_CUDA_STATIC_LINK=0
|
||||
export USE_CUPTI_SO=1
|
||||
fi
|
||||
else
|
||||
echo "Unknown cuda version $CUDA_VERSION"
|
||||
exit 1
|
||||
|
||||
@ -313,7 +313,7 @@ if [[ "$(uname)" == 'Linux' && "$PACKAGE_TYPE" == 'manywheel' ]]; then
|
||||
# Please see issue for reference: https://github.com/pytorch/pytorch/issues/152426
|
||||
if [[ "$(uname -m)" == "s390x" ]]; then
|
||||
cxx_abi="19"
|
||||
elif [[ "$DESIRED_CUDA" != 'cu118' && "$DESIRED_CUDA" != 'xpu' && "$DESIRED_CUDA" != 'rocm'* ]]; then
|
||||
elif [[ "$DESIRED_CUDA" != 'xpu' && "$DESIRED_CUDA" != 'rocm'* ]]; then
|
||||
cxx_abi="18"
|
||||
else
|
||||
cxx_abi="16"
|
||||
|
||||
@ -1,59 +0,0 @@
|
||||
@echo off
|
||||
|
||||
set MODULE_NAME=pytorch
|
||||
|
||||
IF NOT EXIST "setup.py" IF NOT EXIST "%MODULE_NAME%" (
|
||||
call internal\clone.bat
|
||||
cd %~dp0
|
||||
) ELSE (
|
||||
call internal\clean.bat
|
||||
)
|
||||
IF ERRORLEVEL 1 goto :eof
|
||||
|
||||
call internal\check_deps.bat
|
||||
IF ERRORLEVEL 1 goto :eof
|
||||
|
||||
REM Check for optional components
|
||||
|
||||
set USE_CUDA=
|
||||
set CMAKE_GENERATOR=Visual Studio 15 2017 Win64
|
||||
|
||||
IF "%NVTOOLSEXT_PATH%"=="" (
|
||||
IF EXIST "C:\Program Files\NVIDIA Corporation\NvToolsExt\lib\x64\nvToolsExt64_1.lib" (
|
||||
set NVTOOLSEXT_PATH=C:\Program Files\NVIDIA Corporation\NvToolsExt
|
||||
) ELSE (
|
||||
echo NVTX ^(Visual Studio Extension ^for CUDA^) ^not installed, failing
|
||||
exit /b 1
|
||||
)
|
||||
)
|
||||
|
||||
IF "%CUDA_PATH_V118%"=="" (
|
||||
IF EXIST "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin\nvcc.exe" (
|
||||
set "CUDA_PATH_V118=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8"
|
||||
) ELSE (
|
||||
echo CUDA 11.8 not found, failing
|
||||
exit /b 1
|
||||
)
|
||||
)
|
||||
|
||||
IF "%BUILD_VISION%" == "" (
|
||||
set TORCH_CUDA_ARCH_LIST=3.7+PTX;5.0;6.0;6.1;7.0;7.5;8.0;8.6;9.0
|
||||
set TORCH_NVCC_FLAGS=-Xfatbin -compress-all
|
||||
) ELSE (
|
||||
set NVCC_FLAGS=-D__CUDA_NO_HALF_OPERATORS__ --expt-relaxed-constexpr -gencode=arch=compute_35,code=sm_35 -gencode=arch=compute_50,code=sm_50 -gencode=arch=compute_60,code=sm_60 -gencode=arch=compute_70,code=sm_70 -gencode=arch=compute_75,code=sm_75 -gencode=arch=compute_80,code=compute_80 -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_90,code=compute_90
|
||||
)
|
||||
|
||||
set "CUDA_PATH=%CUDA_PATH_V118%"
|
||||
set "PATH=%CUDA_PATH_V118%\bin;%PATH%"
|
||||
|
||||
:optcheck
|
||||
|
||||
call internal\check_opts.bat
|
||||
IF ERRORLEVEL 1 goto :eof
|
||||
|
||||
if exist "%NIGHTLIES_PYTORCH_ROOT%" cd %NIGHTLIES_PYTORCH_ROOT%\..
|
||||
call %~dp0\internal\copy.bat
|
||||
IF ERRORLEVEL 1 goto :eof
|
||||
|
||||
call %~dp0\internal\setup.bat
|
||||
IF ERRORLEVEL 1 goto :eof
|
||||
@ -23,7 +23,6 @@ set CUDNN_LIB_FOLDER="lib\x64"
|
||||
:: Skip all of this if we already have cuda installed
|
||||
if exist "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v%CUDA_VERSION_STR%\bin\nvcc.exe" goto set_cuda_env_vars
|
||||
|
||||
if %CUDA_VER% EQU 118 goto cuda118
|
||||
if %CUDA_VER% EQU 124 goto cuda124
|
||||
if %CUDA_VER% EQU 126 goto cuda126
|
||||
if %CUDA_VER% EQU 128 goto cuda128
|
||||
@ -31,31 +30,6 @@ if %CUDA_VER% EQU 128 goto cuda128
|
||||
echo CUDA %CUDA_VERSION_STR% is not supported
|
||||
exit /b 1
|
||||
|
||||
:cuda118
|
||||
|
||||
set CUDA_INSTALL_EXE=cuda_11.8.0_522.06_windows.exe
|
||||
if not exist "%SRC_DIR%\temp_build\%CUDA_INSTALL_EXE%" (
|
||||
curl -k -L "https://ossci-windows.s3.amazonaws.com/%CUDA_INSTALL_EXE%" --output "%SRC_DIR%\temp_build\%CUDA_INSTALL_EXE%" & REM @lint-ignore
|
||||
if errorlevel 1 exit /b 1
|
||||
set "CUDA_SETUP_FILE=%SRC_DIR%\temp_build\%CUDA_INSTALL_EXE%"
|
||||
set "ARGS=cuda_profiler_api_11.8 thrust_11.8 nvcc_11.8 cuobjdump_11.8 nvprune_11.8 nvprof_11.8 cupti_11.8 cublas_11.8 cublas_dev_11.8 cudart_11.8 cufft_11.8 cufft_dev_11.8 curand_11.8 curand_dev_11.8 cusolver_11.8 cusolver_dev_11.8 cusparse_11.8 cusparse_dev_11.8 npp_11.8 npp_dev_11.8 nvrtc_11.8 nvrtc_dev_11.8 nvml_dev_11.8 nvtx_11.8"
|
||||
)
|
||||
|
||||
set CUDNN_FOLDER=cudnn-windows-x86_64-9.5.0.50_cuda11-archive
|
||||
set CUDNN_LIB_FOLDER="lib"
|
||||
set "CUDNN_INSTALL_ZIP=%CUDNN_FOLDER%.zip"
|
||||
if not exist "%SRC_DIR%\temp_build\%CUDNN_INSTALL_ZIP%" (
|
||||
curl -k -L "http://s3.amazonaws.com/ossci-windows/%CUDNN_INSTALL_ZIP%" --output "%SRC_DIR%\temp_build\%CUDNN_INSTALL_ZIP%" & REM @lint-ignore
|
||||
if errorlevel 1 exit /b 1
|
||||
set "CUDNN_SETUP_FILE=%SRC_DIR%\temp_build\%CUDNN_INSTALL_ZIP%"
|
||||
)
|
||||
|
||||
@REM cuDNN 8.3+ required zlib to be installed on the path
|
||||
echo Installing ZLIB dlls
|
||||
curl -k -L "http://s3.amazonaws.com/ossci-windows/zlib123dllx64.zip" --output "%SRC_DIR%\temp_build\zlib123dllx64.zip"
|
||||
7z x "%SRC_DIR%\temp_build\zlib123dllx64.zip" -o"%SRC_DIR%\temp_build\zlib"
|
||||
xcopy /Y "%SRC_DIR%\temp_build\zlib\dll_x64\*.dll" "C:\Windows\System32"
|
||||
|
||||
goto cuda_common
|
||||
|
||||
:cuda126
|
||||
|
||||
14
.github/scripts/generate_binary_build_matrix.py
vendored
14
.github/scripts/generate_binary_build_matrix.py
vendored
@ -40,19 +40,6 @@ CUDA_AARCH64_ARCHES = ["12.8-aarch64"]
|
||||
|
||||
|
||||
PYTORCH_EXTRA_INSTALL_REQUIREMENTS = {
|
||||
"11.8": (
|
||||
"nvidia-cuda-nvrtc-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | " # noqa: B950
|
||||
"nvidia-cuda-runtime-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cuda-cupti-cu11==11.8.87; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cudnn-cu11==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cublas-cu11==11.11.3.6; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cufft-cu11==10.9.0.58; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-curand-cu11==10.3.0.86; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cusolver-cu11==11.4.1.48; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cusparse-cu11==11.7.5.86; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-nccl-cu11==2.21.5; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-nvtx-cu11==11.8.86; platform_system == 'Linux' and platform_machine == 'x86_64'"
|
||||
),
|
||||
"12.6": (
|
||||
"nvidia-cuda-nvrtc-cu12==12.6.77; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
"nvidia-cuda-runtime-cu12==12.6.77; platform_system == 'Linux' and platform_machine == 'x86_64' | "
|
||||
@ -413,4 +400,3 @@ def generate_wheels_matrix(
|
||||
|
||||
validate_nccl_dep_consistency("12.8")
|
||||
validate_nccl_dep_consistency("12.6")
|
||||
validate_nccl_dep_consistency("11.8")
|
||||
|
||||
2
.github/workflows/build-almalinux-images.yml
vendored
2
.github/workflows/build-almalinux-images.yml
vendored
@ -36,7 +36,7 @@ jobs:
|
||||
runs-on: linux.9xlarge.ephemeral
|
||||
strategy:
|
||||
matrix:
|
||||
tag: ["cuda11.8", "cuda12.6", "cuda12.8", "cuda12.9", "rocm6.3", "rocm6.4", "cpu"]
|
||||
tag: ["cuda12.6", "cuda12.8", "cuda12.9", "rocm6.3", "rocm6.4", "cpu"]
|
||||
steps:
|
||||
- name: Build docker image
|
||||
uses: pytorch/pytorch/.github/actions/binary-docker-build@main
|
||||
|
||||
1
.github/workflows/build-libtorch-images.yml
vendored
1
.github/workflows/build-libtorch-images.yml
vendored
@ -50,7 +50,6 @@ jobs:
|
||||
include: [
|
||||
{ tag: "cuda12.8" },
|
||||
{ tag: "cuda12.6" },
|
||||
{ tag: "cuda11.8" },
|
||||
{ tag: "rocm6.3" },
|
||||
{ tag: "rocm6.4" },
|
||||
{ tag: "cpu" },
|
||||
|
||||
2
.github/workflows/build-magma-linux.yml
vendored
2
.github/workflows/build-magma-linux.yml
vendored
@ -34,7 +34,7 @@ jobs:
|
||||
id-token: write
|
||||
strategy:
|
||||
matrix:
|
||||
cuda_version: ["129", "128", "126", "118"]
|
||||
cuda_version: ["129", "128", "126"]
|
||||
steps:
|
||||
- name: Checkout PyTorch
|
||||
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
|
||||
|
||||
1
.github/workflows/build-manywheel-images.yml
vendored
1
.github/workflows/build-manywheel-images.yml
vendored
@ -49,7 +49,6 @@ jobs:
|
||||
include: [
|
||||
{ name: "manylinux2_28-builder", tag: "cuda12.8", runner: "linux.9xlarge.ephemeral" },
|
||||
{ name: "manylinux2_28-builder", tag: "cuda12.6", runner: "linux.9xlarge.ephemeral" },
|
||||
{ name: "manylinux2_28-builder", tag: "cuda11.8", runner: "linux.9xlarge.ephemeral" },
|
||||
{ name: "manylinuxaarch64-builder", tag: "cuda12.8", runner: "linux.arm64.2xlarge.ephemeral" },
|
||||
{ name: "manylinux2_28-builder", tag: "rocm6.3", runner: "linux.9xlarge.ephemeral" },
|
||||
{ name: "manylinux2_28-builder", tag: "rocm6.4", runner: "linux.9xlarge.ephemeral" },
|
||||
|
||||
6
.github/workflows/test-check-binary.yml
vendored
6
.github/workflows/test-check-binary.yml
vendored
@ -34,7 +34,9 @@ jobs:
|
||||
docker-image: python:3.11
|
||||
docker-build-dir: "skip-docker-build"
|
||||
script: |
|
||||
STABLE_CUDA_VERSION=$(python3 .github/scripts/get_ci_variable.py --cuda-stable-version)
|
||||
CUDA_VERSION_NODOT=$(echo ${STABLE_CUDA_VERSION} | tr -d '.')
|
||||
pushd .ci/pytorch/
|
||||
pip install --pre torch --index-url https://download.pytorch.org/whl/nightly/cu124
|
||||
DESIRED_PYTHON=3.11 DESIRED_CUDA=cu124 PACKAGE_TYPE=manywheel ./check_binary.sh
|
||||
pip install --pre torch --index-url https://download.pytorch.org/whl/nightly/cu${CUDA_VERSION_NODOT}
|
||||
DESIRED_PYTHON=3.11 DESIRED_CUDA=cu${CUDA_VERSION_NODOT} PACKAGE_TYPE=manywheel ./check_binary.sh
|
||||
popd
|
||||
|
||||
@ -53,7 +53,6 @@ from torch.profiler._pattern_matcher import (
|
||||
SynchronizedDataLoaderPattern,
|
||||
)
|
||||
from torch.testing._internal.common_cuda import TEST_MULTIGPU
|
||||
from torch.testing._internal.common_device_type import skipCUDAVersionIn
|
||||
from torch.testing._internal.common_utils import (
|
||||
instantiate_parametrized_tests,
|
||||
IS_ARM64,
|
||||
@ -102,7 +101,6 @@ except ModuleNotFoundError:
|
||||
@unittest.skipIf(IS_WINDOWS, "Test is flaky on Windows")
|
||||
@unittest.skipIf(not torch.cuda.is_available(), "CUDA is required")
|
||||
class TestProfilerCUDA(TestCase):
|
||||
@skipCUDAVersionIn([(11, 5)]) # https://github.com/pytorch/pytorch/issues/69023
|
||||
def test_mem_leak(self):
|
||||
"""Checks that there's no memory leak when using profiler with CUDA"""
|
||||
t = torch.rand(1, 1).cuda()
|
||||
|
||||
@ -29,7 +29,7 @@ from torch.testing._internal.common_device_type import \
|
||||
(instantiate_device_type_tests, dtypes, has_cusolver, has_hipsolver,
|
||||
onlyCPU, skipCUDAIf, skipCUDAIfNoMagma, skipCPUIfNoLapack, precisionOverride,
|
||||
skipCUDAIfNoMagmaAndNoCusolver, skipCUDAIfRocm, onlyNativeDeviceTypes, dtypesIfCUDA,
|
||||
onlyCUDA, skipCUDAVersionIn, skipMeta, skipCUDAIfNoCusolver, skipCUDAIfNotRocm, skipCUDAIfRocmVersionLessThan,
|
||||
onlyCUDA, skipMeta, skipCUDAIfNoCusolver, skipCUDAIfNotRocm, skipCUDAIfRocmVersionLessThan,
|
||||
dtypesIfMPS, largeTensorTest)
|
||||
from torch.testing import make_tensor
|
||||
from torch.testing._internal.common_dtype import (
|
||||
@ -3713,7 +3713,6 @@ class TestLinalg(TestCase):
|
||||
|
||||
@skipCUDAIfNoMagma
|
||||
@skipCPUIfNoLapack
|
||||
@skipCUDAVersionIn([(11, 6), (11, 7)]) # https://github.com/pytorch/pytorch/issues/75391
|
||||
@dtypes(*floating_and_complex_types())
|
||||
def test_matrix_rank_empty(self, device, dtype):
|
||||
matrix_rank = torch.linalg.matrix_rank
|
||||
|
||||
@ -16,8 +16,7 @@ from torch.testing._internal.common_dtype import (
|
||||
floating_types, floating_and_complex_types_and, get_all_fp_dtypes)
|
||||
from torch.testing._internal.common_device_type import (
|
||||
_TestParametrizer, _update_param_kwargs, expectedFailureMPS, toleranceOverride, tol,
|
||||
skipCUDAIfRocm, precisionOverride, skipMeta, skipMPS,
|
||||
skipCUDAVersionIn)
|
||||
skipCUDAIfRocm, precisionOverride, skipMeta, skipMPS)
|
||||
from torch.testing._internal.common_methods_invocations import DecorateInfo
|
||||
from torch.testing._internal.common_nn import (
|
||||
cosineembeddingloss_reference, cross_entropy_loss_reference, ctcloss_reference,
|
||||
@ -3172,14 +3171,6 @@ rnn_gru_lstm_module_info_decorators = (
|
||||
DecorateInfo(
|
||||
unittest.expectedFailure, "TestModule", "test_non_contiguous_tensors",
|
||||
active_if=(TEST_CUDNN and TEST_WITH_ROCM), dtypes=(torch.float,), device_type='cuda'
|
||||
),
|
||||
DecorateInfo(
|
||||
skipCUDAVersionIn([(11, 7)]), "TestExpandedWeightModule", "test_module",
|
||||
device_type='cuda'
|
||||
),
|
||||
DecorateInfo(
|
||||
skipCUDAVersionIn([(11, 7)]), "TestDecomp", "test_rnn_decomp_module",
|
||||
device_type='cuda'
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user