Build error: C++ compiler produced errors. See the Build Log for further details

Hi! I ask a question here! I getting wrong on gpu coder says "Build error: C++ compiler produced errors. See the Build Log for further details" but i already installed mingw and i also replace attachment of bug report how can i fix this guys Thanks

[Joss Knight]

What does the Build Log say?

[Allan III T. Condiman]

Cannot find cl.exe in Path

Answer 1

[Walter Roberson]

You must install Visual Studio to use GPU Coder.

[Allan III T. Condiman]

What Visual Studio Community Sir?

[Walter Roberson]

Any one that is compatible with your release. I seem to recall that vs 2019 Community is supported for your release.

[Allan III T. Condiman]

Thank you very much sir :)

[Ali Al-Saegh]

Hello,

I have the same proble as Allan stated. I installed VS 2019 Community but still have the same problem. Are there any configuration should me made to solve the problem.

I need help please.

Answer 2

[Allan III T. Condiman]

Cannot find cl.exe in Path

[Joss Knight]

It also says, right there: "'MINGW64 Compiler (C++)' is not supported for GPU code generation."

[Ultimate CR]

Hi @Joss Knight, I am using MATLAB 2019b version and have installed VS 2019 community version. I already have built MKLDNN library successfully but still getting the same error as of now.

Error is, MEX build failed: Build error: C++ compiler produced errors. See the Build Log for further details. <a href="matlab:helpview('C:/Program Files/Polyspace/R2019b/help/coder/helptargets.map','msginfo_Coder_FE_TargetBuildError');"

Solutions I already have tried are:

1. https://www.mathworks.com/matlabcentral/answers/761406-code-generation-error-build-error-c-compiler-produced-errors-see-the-build-log-for-further-deta which talks about placing mkldnn_config.h into mex file
2. https://www.mathworks.com/matlabcentral/answers/447387-matlab-coder-how-do-i-build-the-intel-mkl-dnn-library-for-deep-learning-c-code-generation-and-dep which talks about build mkl-dnn from source where I got the sucess.
3. mex -setup c++ link that matlab central provides.

I am working on a research project where I want to see the inference time by resnet50 model using resnet50.onnx model. My Question is,

Do we need the MEX build to perform inference from the generated c++ code? Because error is not coming in code generation.

If yes, then how to debug this error.

@Allan Condiman, @Walter Roberson and @Ali Al-Saegh are you able to solve this issue? If yes, please help.

[Joss Knight]

This is a new question, not really a comment on the answer to an existing question. Try posting a new question, which might draw the attention of the GPU Coder team. That said:

• Read the documentation for GPU code generation
• Try coder.checkGpuInstall or the GPU Environment Check and Setup App. It might help work out why Visual Studio is not on the path.
• MKL-DNN is for CPU execution, so either you don't have a GPU or you are building with the wrong target
• You don't need to do code generation to run resnet50 in MATLAB, for instance, try:

net = resnet50;
x = imresize(im2single(imread('peppers.png')),net.Layers(1).InputSize(1:2));
predict(net,x)

ans = 1×1000

1.0e+00 * 0.0001 0.0008 0.0002 0.0002 0.0003 0.0001 0.0001 0.0002 0.0001 0.0004 0.0002 0.0002 0.0001 0.0001 0.0000 0.0002 0.0001 0.0001 0.0001 0.0001 0.0003 0.0003 0.0004 0.0001 0.0003 0.0003 0.0001 0.0001 0.0007 0.0005