nvcc fatal : 'gpu-architecture'옵션에 'sm_20'값이 정의되어 있지 않습니다.

Question

많은 페이지를 보았지만 불분명하고 지식이 충분하지 않아서 말하는 내용을 따르지 못했습니다. .

나는 실행하려고 :

luarocks install https://raw.githubusercontent.com/qassemoquab/stnbhwd/master/stnbhwd-scm-1.rockspec

을 그래서 GPU 가속을 사용하여 일부 이미지를 통해 DenseCap를 실행할 수있다. 내가 그것을 실행하면이 오류를 얻을 :

$ luarocks install https://raw.githubusercontent.com/qassemoquab/stnbhwd/master/stnbhwd-scm-1.rockspec 
Using https://raw.githubusercontent.com/qassemoquab/stnbhwd/master/stnbhwd-scm-1.rockspec... switching to 'build' mode 
Cloning into 'stnbhwd'... 
remote: Counting objects: 24, done. 
remote: Compressing objects: 100% (23/23), done. 
remote: Total 24 (delta 0), reused 14 (delta 0), pack-reused 0 
Receiving objects: 100% (24/24), 19.42 KiB | 0 bytes/s, done. 
Checking connectivity... done. 
cmake -E make_directory build && cd build && cmake .. -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH="/home/tex/torch/install/bin/.." -DCMAKE_INSTALL_PREFIX="/home/tex/torch/install/lib/luarocks/rocks/stnbhwd/scm-1" && make 

-- The C compiler identification is GNU 5.4.0 
-- The CXX compiler identification is GNU 5.4.0 
-- Check for working C compiler: /usr/bin/cc 
-- Check for working C compiler: /usr/bin/cc -- works 
-- Detecting C compiler ABI info 
-- Detecting C compiler ABI info - done 
-- Detecting C compile features 
-- Detecting C compile features - done 
-- Check for working CXX compiler: /usr/bin/c++ 
-- Check for working CXX compiler: /usr/bin/c++ -- works 
-- Detecting CXX compiler ABI info 
-- Detecting CXX compiler ABI info - done 
-- Detecting CXX compile features 
-- Detecting CXX compile features - done 
-- Found Torch7 in /home/tex/torch/install 
-- Try OpenMP C flag = [-fopenmp] 
-- Performing Test OpenMP_FLAG_DETECTED 
-- Performing Test OpenMP_FLAG_DETECTED - Success 
-- Try OpenMP CXX flag = [-fopenmp] 
-- Performing Test OpenMP_FLAG_DETECTED 
-- Performing Test OpenMP_FLAG_DETECTED - Success 
-- Found OpenMP: -fopenmp 
-- Compiling with OpenMP support 
-- Looking for pthread.h 
-- Looking for pthread.h - found 
-- Looking for pthread_create 
-- Looking for pthread_create - not found 
-- Looking for pthread_create in pthreads 
-- Looking for pthread_create in pthreads - not found 
-- Looking for pthread_create in pthread 
-- Looking for pthread_create in pthread - found 
-- Found Threads: TRUE 
-- Found CUDA: /usr/local/cuda (found suitable version "9.0", minimum required is "5.5") 
-- Configuring done 
-- Generating done 
-- Build files have been written to: /tmp/luarocks_stnbhwd-scm-1-4197/stnbhwd/build 
Scanning dependencies of target stn 
[ 25%] Building C object CMakeFiles/stn.dir/init.c.o 
[ 50%] Linking C shared module libstn.so 
[ 50%] Built target stn 
[ 75%] Building NVCC (Device) object CMakeFiles/custn.dir/custn_generated_init.cu.o 
nvcc fatal : Value 'sm_20' is not defined for option 'gpu-architecture' 
CMake Error at custn_generated_init.cu.o.cmake:207 (message): 
    Error generating 
    /tmp/luarocks_stnbhwd-scm-1-4197/stnbhwd/build/CMakeFiles/custn.dir//./custn_generated_init.cu.o 


CMakeFiles/custn.dir/build.make:63: recipe for target 'CMakeFiles/custn.dir/custn_generated_init.cu.o' failed 
make[2]: *** [CMakeFiles/custn.dir/custn_generated_init.cu.o] Error 1 
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/custn.dir/all' failed 
make[1]: *** [CMakeFiles/custn.dir/all] Error 2 
Makefile:127: recipe for target 'all' failed 
make: *** [all] Error 2 

Error: Build error: Failed building. 

내가 luarocks install cutorch, luarocks install cunn 및 luarocks install cudnn 잘 할 수 있었다.

저는 GTX 1080ti에서 우분투 16.04를 사용하고 있습니다. 내가 $ nvcc -V을 실행할 때 몇 가지 이유를 들어

$ nvidia-smi 
Tue Dec 5 16:25:42 2017  
+-----------------------------------------------------------------------------+ 
| NVIDIA-SMI 384.90     Driver Version: 384.90     | 
|-------------------------------+----------------------+----------------------+ 
| GPU Name  Persistence-M| Bus-Id  Disp.A | Volatile Uncorr. ECC | 
| Fan Temp Perf Pwr:Usage/Cap|   Memory-Usage | GPU-Util Compute M. | 
|===============================+======================+======================| 
| 0 GeForce GTX 108... Off | 00000000:29:00.0 On |     N/A | 
| 0% 47C P8 16W/250W | 716MiB/11169MiB |  1%  Default | 
+-------------------------------+----------------------+----------------------+ 

+-----------------------------------------------------------------------------+ 
| Processes:              GPU Memory | 
| GPU  PID Type Process name        Usage  | 
|=============================================================================| 
| 0  1128  G /usr/lib/xorg/Xorg       479MiB | 
| 0  1782  G compiz          234MiB | 
+-----------------------------------------------------------------------------+ 

는, 내가 얻을 :

The program 'nvcc' is currently not installed. You can install it by typing: 
sudo apt install nvidia-cuda-toolkit 

이 .. 그리고 이것이 내가 직면하고 또 다른 큰 문제입니다. nvcc를 설치할 때 Cuda 7.5 용 툴킷을 설치하지만 Cuda 9.0이 있습니다. Cuda's website에서 오는 .deb 파일을 사용하여 설치했습니다.

$ sudo apt-get install cuda 
Reading package lists... Done 
Building dependency tree  
Reading state information... Done 
cuda is already the newest version (9.0.176-1). 
The following packages were automatically installed and are no longer required: 
    libcublas7.5 libcudart7.5 libcufft7.5 libcufftw7.5 libcuinj64-7.5 
    libcurand7.5 libcusolver7.5 libcusparse7.5 libnppc7.5 libnppi7.5 libnpps7.5 
    libnvblas7.5 libnvrtc7.5 libnvtoolsext1 libnvvm3 libthrust-dev libvdpau-dev 
    nvidia-cuda-dev nvidia-cuda-doc nvidia-cuda-gdb nvidia-opencl-dev 
    nvidia-profiler nvidia-visual-profiler opencl-headers 
Use 'sudo apt autoremove' to remove them. 
0 upgraded, 0 newly installed, 0 to remove and 222 not upgraded. 

솔직히 내가 NVCC를 설치해야 할 것으로 예상하지만, 내가 말했듯이, 그것은 잘못된 버전을 설치하고, 나는 내 인생에 맞는 버전을 설치하는 방법을 알아낼 수 없습니다, 그리고 임 정말 내가 왜이 오류가 발생하는지에 관해 혼란스러워합니다. 이미 cutorch, cudnn 및 cunn을 설치 했는데도 문제가 없습니다. 어떤 도움을 설치하려고하는 stnbhwd의 CMakeLists.txt에서 일부 상위 버전에 코드 아키텍처 (예 : sm_20를) 변경 ...

감사

Answer 1

시도를 감사합니다.

IF (CUDA_FOUND) 
    LIST(APPEND CUDA_NVCC_FLAGS "-arch=sm_20") 

에 :에서

IF (CUDA_FOUND) 
    LIST(APPEND CUDA_NVCC_FLAGS "-arch=sm_30")