error in vl_compilenn('enableGpu',true)

[supreet21]

I am using Cuda toolkit 10.2 and MatLab 2019b with visual studio 2019 compiler as well as parallel computing toolbox. The error was generated after I run the command above.

Error using vl_compilenn>search_cuda_devkit (line 715)
Could not find a valid NVCC executable\n

Error in vl_compilenn (line 279)
if isempty(opts.cudaRoot), opts.cudaRoot = search_cuda_devkit(opts) ; end

Any help and guidance will be much appreciated. Thank you in advance.

[jotaf98]

Hi, you need to locate the nvcc executable in your system and pass that folder in as an extra option: vl_compilenn(..., 'cudaRoot', '<the_path>')

If that option is not given then search_cuda_devkit will look for it in a few usual places (though it's not guaranteed as many systems are different).

[supreet21]

thanks

> vl_compilenn('enableGpu', true, 'cudaRoot', 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2')
this command solved the problem

[supreet21]

But there is another issue.

Warning: Support for toolboxdir('distcomp') will be removed in a future release. Use toolboxdir('parallel') instead.

In toolboxdir (line 49)
In vl_compilenn (line 367)
Warning: CL.EXE not found in PATH. Trying to guess out of mex setup.
In vl_compilenn>check_clpath (line 650)
In vl_compilenn (line 426)
'cl.exe' is not recognized as an internal or external command,
operable program or batch file.
Error using vl_compilenn>check_clpath (line 656)
Unable to find cl.exe

Error in vl_compilenn (line 426)
cl_path = fileparts(check_clpath()); % check whether cl.exe in path

then solved the problem by referring to the following link.
https://stackoverflow.com/questions/40226354/matconvnet-error-cl-exe-not-found

changes made were

1.for 1st trail
cl_path = 'C:\Program Files (x86)\Microsoft Visual
Studio\2019\Community\VC\Tools\MSVC\14.25.28610\bin\Hostx64\x64' after line(646)
2. And in the 2nd trail

addpath(fullfile('C:', 'Program Files (x86)', 'Microsoft Visual Studio', '2017', 'Professional', 'VC', 'Tools', 'MSVC', '14.16.27023', 'bin', 'Hostx64', 'x64'));

[supreet21]

But following this we encounter with another issue

on execution of the command

vl_compilenn('enableGpu', true, 'cudaRoot', 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2')
or
vl_compilenn('enableGpu', true, 'cudaRoot', 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2','EnableImreadJpeg', false)

Warning: Support for toolboxdir('distcomp') will be removed in a future release. Use toolboxdir('parallel') instead.

In toolboxdir (line 49)
In vl_compilenn (line 367)
nvcc fatal : '-DNDEBUG': expected a number
Error using vl_compilenn>nvcc_compile (line 615)
Command "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\bin\nvcc" -c -o
"C:\Users\Hp\AppData\Roaming\MathWorks\MATLAB
Add-Ons\Collections\vlfeat_matconvnet\vlfeat-matconvnet-1e5ae7b\matlab\mex.build\bits\data.obj"
"C:\Users\Hp\AppData\Roaming\MathWorks\MATLAB
Add-Ons\Collections\vlfeat_matconvnet\vlfeat-matconvnet-1e5ae7b\matlab\src\bits\data.cu" -DENABLE_GPU -DENABLE_DOUBLE -O
-DNDEBUG -D_FORCE_INLINES --std=c++11 -I"C:\Program Files\MATLAB\R2019b\extern\include" -I"C:\Program
Files\MATLAB\R2019b\toolbox\parallel\gpu\extern\include" -gencode=arch=compute_50,code="sm_50,compute_50"
--compiler-options=/MD
--compiler-bindir="C:\ProgramData\MATLAB\SupportPackages\R2019b\3P.instrset\mingw_w64.instrset\VC\bin" failed.

Error in vl_compilenn (line 487)
nvcc_compile(opts, srcs{i}, objfile, flags) ;

[vishwa91]

You may want to try the following:

1. Comment flags.base{end+1} = '-O0' on line 340
2. Switch to mex_cuda -- vl_compilenn('verbose', 1, 'enableGpu', true, 'cudaRoot', 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2', 'CudaMethod', 'mex')

This worked for me on Windows 10.

[NikolaiSkuratov]

Thank you very much, vishwa91! I was about to give up, but your ssuggestion helped me to work out the problem.