![[PukiWiki]](image/pukiwiki.png "[PukiWiki]")
|
Python CUDA on MacBookPro |
| macos | Xcode | Apple LLVM | |
| CUDA Toolkit 9.1 対応 | 10.13.3 | 9.2 | 9.0 |
| 最新版 | 10.13.4 | 9.3 | 9.1 |
[/Developer/NVIDIA/CUDA-9.1/samples]$ make -C 0_Simple/vectorAdd
/Developer/NVIDIA/CUDA-9.1/bin/nvcc -ccbin clang++ -I../../common/inc -m64 -Xcompiler -arch -Xcompiler x86_64 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_70,code=compute_70 -o vectorAdd.o -c vectorAdd.cu
nvcc fatal : The version ('90100') of the host compiler ('Apple clang') is not supported
make: *** [vectorAdd.o] Error 1
$ mv ~/Downloads/Xcode.app /Applications/Xcode_9.2.app
[/Applications]$ clang --version
Apple LLVM version 9.1.0 (clang-902.0.39.1)
Target: x86_64-apple-darwin17.5.0
Thread model: posix
InstalledDir: /Library/Developer/CommandLineTools/usr/bin
[/Applications]$ sudo xcode-select -s /Applications/Xcode_9.2.app/Contents/Developer/
[/Applications]$ xcodebuild -version
Xcode 9.2
Build version 9C40b
[/Applications]$ clang --version
Apple LLVM version 9.0.0 (clang-900.0.39.2)
Target: x86_64-apple-darwin17.5.0
Thread model: posix
InstalledDir: /Applications/Xcode_9.2.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin
export DEVELOPER_DIR=/Applications/Xcode_9.2.app/Contents/Developer/
[/Developer/NVIDIA/CUDA-9.1/samples]$ ./bin/x86_64/darwin/release/deviceQuery
./bin/x86_64/darwin/release/deviceQuery Starting...
CUDA Device Query (Runtime API) version (CUDART static linking)
cudaGetDeviceCount returned 35
-> CUDA driver version is insufficient for CUDA runtime version
Result = FAIL
export PS1="[\w]\$ "
export PYENV_ROOT="${HOME}/.pyenv"
export PATH=${PYENV_ROOT}/bin:$PATH
eval "$(pyenv init -)"
PERL_MB_OPT="--install_base \"/Users/atsushi/perl5\""; export PERL_MB_OPT;
PERL_MM_OPT="INSTALL_BASE=/Users/atsushi/perl5"; export PERL_MM_OPT;
export JAVA_HOME="/Library/Java/JavaVirtualMachines/jdk1.8.0_92.jdk/Contents/Home/jre/"
export PATH=${PATH}:${JAVA_HOME}/bin
export SCALA_HOME=/usr/local/Cellar/scala/2.11.4/
export LANG="ja_JP.UTF-8"
export ANSIBLE_NOCOWS=1
# for CUDA
export DEVELOPER_DIR=/Applications/Xcode_9.2.app/Contents/Developer/
export PATH=/Developer/NVIDIA/CUDA-9.1/bin${PATH:+:${PATH}}
export DYLD_LIBRARY_PATH=/Developer/NVIDIA/CUDA-9.1/lib${DYLD_LIBRARY_PATH:+:${DYLD_LIBRARY_PATH}
[/Developer/NVIDIA/CUDA-9.1/samples]$ sudo make -C 1_Utilities/deviceQuery
[/Developer/NVIDIA/CUDA-9.1/samples]$ ./bin/x86_64/darwin/release/deviceQuery
./bin/x86_64/darwin/release/deviceQuery Starting...
CUDA Device Query (Runtime API) version (CUDART static linking)
Detected 1 CUDA Capable device(s)
Device 0: "GeForce GT 750M"
CUDA Driver Version / Runtime Version 9.1 / 9.1
CUDA Capability Major/Minor version number: 3.0
Total amount of global memory: 2048 MBytes (2147024896 bytes)
( 2) Multiprocessors, (192) CUDA Cores/MP: 384 CUDA Cores
GPU Max Clock rate: 926 MHz (0.93 GHz)
Memory Clock rate: 2508 Mhz
Memory Bus Width: 128-bit
L2 Cache Size: 262144 bytes
Maximum Texture Dimension Size (x,y,z) 1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)
Maximum Layered 1D Texture Size, (num) layers 1D=(16384), 2048 layers
Maximum Layered 2D Texture Size, (num) layers 2D=(16384, 16384), 2048 layers
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 49152 bytes
Total number of registers available per block: 65536
Warp size: 32
Maximum number of threads per multiprocessor: 2048
Maximum number of threads per block: 1024
Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
Max dimension size of a grid size (x,y,z): (2147483647, 65535, 65535)
Maximum memory pitch: 2147483647 bytes
Texture alignment: 512 bytes
Concurrent copy and kernel execution: Yes with 1 copy engine(s)
Run time limit on kernels: Yes
Integrated GPU sharing Host Memory: No
Support host page-locked memory mapping: Yes
Alignment requirement for Surfaces: Yes
Device has ECC support: Disabled
Device supports Unified Addressing (UVA): Yes
Supports Cooperative Kernel Launch: No
Supports MultiDevice Co-op Kernel Launch: No
Device PCI Domain ID / Bus ID / location ID: 0 / 1 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 9.1, CUDA Runtime Version = 9.1, NumDevs = 1
Result = PASS
[~/Downloads]$ tar xzvf cudnn-9.1-osx-x64-v7-ga.tgz
x cuda/include/cudnn.h
x cuda/NVIDIA_SLA_cuDNN_Support.txt
x cuda/lib/libcudnn.7.dylib
x cuda/lib/libcudnn.dylib
x cuda/lib/libcudnn_static.a
[~/Downloads]$ sudo cp -v cuda/include/cudnn.h /usr/local/cuda/include/
cuda/include/cudnn.h -> /usr/local/cuda/include/cudnn.h
[~/Downloads]$ sudo cp -v cuda/lib/libcudnn* /usr/local/cuda/lib
cuda/lib/libcudnn.7.dylib -> /usr/local/cuda/lib/libcudnn.7.dylib
cuda/lib/libcudnn.dylib -> /usr/local/cuda/lib/libcudnn.dylib
cuda/lib/libcudnn_static.a -> /usr/local/cuda/lib/libcudnn_static.a
[~/Downloads]$ sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib/libcudnn*
export PS1="[\w]\$ "
export LANG="ja_JP.UTF-8"
# for pyenv
export PYENV_ROOT="${HOME}/.pyenv"
export PATH=${PYENV_ROOT}/bin:$PATH
eval "$(pyenv init -)"
# for anaconda
. /Users/atsushi/.pyenv/versions/anaconda3-5.1.0/etc/profile.d/conda.sh
PERL_MB_OPT="--install_base \"/Users/atsushi/perl5\""; export PERL_MB_OPT;
PERL_MM_OPT="INSTALL_BASE=/Users/atsushi/perl5"; export PERL_MM_OPT;
export JAVA_HOME="/Library/Java/JavaVirtualMachines/jdk1.8.0_92.jdk/Contents/Home/jre/"
export PATH=${PATH}:${JAVA_HOME}/bin
export SCALA_HOME=/usr/local/Cellar/scala/2.11.4/
export ANSIBLE_NOCOWS=1
# for CUDA
export DEVELOPER_DIR=/Applications/Xcode_9.2.app/Contents/Developer/
export PATH=/Developer/NVIDIA/CUDA-9.1/bin${PATH:+:${PATH}}
export DYLD_LIBRARY_PATH=/Developer/NVIDIA/CUDA-9.1/lib${DYLD_LIBRARY_PATH:+:${DYLD_LIBRARY_PATH}}
# for cuDNN
export DYLD_LIBRARY_PATH=/usr/local/cuda/lib:$DYLD_LIBRARY_PATH
$ echo -e '#include"cudnn.h"\n int main(){}' | nvcc -x c - -o /dev/null \
-Iusr/local/cuda/include -L/usr/local/cuda/lib -lcudnn
これでエラーが起きなきゃ、nvcc と ヘッダファイルがきちんと配布されていることが検証できる
gpu.png