Running examples in Windows.

[xistocarlos]

Hi again,

I wonder if anyone has tried to run the examples on windows recently. I have managed to compile and install the library as well as compile example 1_.

I get the following while linking:

[ 50%] Building Fortran object CMakeFiles/simplenet_infer_fortran.dir/simplenet_infer_fortran.f90.obj
[100%] Linking Fortran executable simplenet_infer_fortran.exe
LINK: command "C:\PROGRA~2\Intel\oneAPI\compiler\latest\bin\ifx.exe /nologo @CMakeFiles\simplenet_infer_fortran.dir\objects1.rsp /Qoption,link,/machine:x64 /Qoption,link,/INCREMENTAL:NO /Qoption,link,/subsystem:console C:\Program Files (x86)\FTorch\lib\ftorch.lib user32.lib /link /out:simplenet_infer_fortran.exe /implib:simplenet_infer_fortran.lib /pdb:C:\Users\xisto\Documents\GitHub\gestpan_working_dir\FTorch-main\examples\1_SimpleNet\build\simplenet_infer_fortran.pdb /version:0.0 /MANIFEST:EMBED,ID=1" failed (exit code 1120) with the following output:
simplenet_infer_fortran.f90.obj : error LNK2019: unresolved external symbol FTORCH_mp_TORCH_TENSOR_FROM_ARRAY_REAL32_1D referenced in function MAIN__
simplenet_infer_fortran.f90.obj : error LNK2019: unresolved external symbol FTORCH_mp_TORCH_MODEL_LOAD referenced in function MAIN__
simplenet_infer_fortran.f90.obj : error LNK2019: unresolved external symbol FTORCH_mp_TORCH_MODEL_FORWARD referenced in function MAIN__
simplenet_infer_fortran.f90.obj : error LNK2019: unresolved external symbol FTORCH_TEST_UTILS_mp_ASSERT_ALLCLOSE_REAL32_1D referenced in function MAIN__
simplenet_infer_fortran.f90.obj : error LNK2019: unresolved external symbol FTORCH_mp_TORCH_MODEL_DELETE referenced in function MAIN__
simplenet_infer_fortran.f90.obj : error LNK2019: unresolved external symbol FTORCH_mp_TORCH_TENSOR_ARRAY_DELETE referenced in function MAIN__
simplenet_infer_fortran.exe : fatal error LNK1120: 6 unresolved externals
NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio\2022\Community\Common7\IDE\CommonExtensions\Microsoft\CMake\CMake\bin\cmake.exe" -E vs_link_exe --intdir=CMakeFiles\simplenet_infer_fortran.dir --rc=C:\PROGRA~2\WI3CF2~1\10\bin\100226~1.0\x64\rc.exe --mt=C:\PROGRA~2\WI3CF2~1\10\bin\100226~1.0\x64\mt.exe --manifests -- C:\PROGRA~2\Intel\oneAPI\compiler\latest\bin\ifx.exe /nologo @CMakeFiles\simplenet_infer_fortran.dir\objects1.rsp @C:\Users\xisto\AppData\Local\Temp\nm97E9.tmp' : return code '0xffffffff'
Stop.
NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.42.34433\bin\HostX64\x64\nmake.exe" -s -f CMakeFiles\simplenet_infer_fortran.dir\build.make /nologo -SL                 CMakeFiles\simplenet_infer_fortran.dir\build' : return code '0x2'
Stop.
NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.42.34433\bin\HostX64\x64\nmake.exe" -s -f CMakeFiles\Makefile2 /nologo -LS                 all' : return code '0x2'
Stop.

Any ideas?

[jatkinson1000]

Hi @xistocarlos
We have built and run the examples on windows in the past, but I will admit that is not our main focus or area of expertise.

It seems that it is not finding the FTorch library to link to.

I believe this is similar to the issue that was raised in #124 so perhaps you can read through that and see if you can follow the same solution?

Feel free to ask further questions here, however.
The fixes related to the above issue are dependent on those discussed in your previous one and have taken some time to resolve with the folks over at pytorch (see linked issues from the above 124).
We are hoping that we will be able to merge quite a few PRs to resolve Windows problems in the near future.

[jatkinson1000]

Amazing, thank you for the update, and thanks @TomMelt for your input.

@xistocarlos if you would be so kind would you mind adding the steps you took to build and run in case any future users come along, thanks?

[xistocarlos]

Steps to make it work with oneAPI Base Toolkit 2025, pytorch 2.4.1. and using the long-long-fix branch.

1. Follow the steps proposed by Tom here:

call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"

cd C:\Users\Quickemu\Documents\FTorch\src
rmdir /s /q build

cd C:\Users\Quickemu\Documents\FTorch\src
cmake -Bbuild -G "NMake Makefiles" -DCMAKE_Fortran_FLAGS="/fpscomp:logicals" -DCMAKE_PREFIX_PATH="C:\Users\Quickemu\Documents\FTorch\.ftorch\Lib\site-packages" -DCMAKE_BUILD_TYPE=Release -DCMAKE_Fortran_COMPILER=ifx -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icx
cmake --build build
cmake --install build

cd C:\Users\Quickemu\Documents\FTorch\examples\1_SimpleNet
rmdir /s /q build
cmake -Bbuild -G "NMake Makefiles" -DCMAKE_PREFIX_PATH="C:/Program Files (x86)/FTorch/lib" -DCMAKE_BUILD_TYPE=Release -DCMAKE_Fortran_COMPILER=ifx
cmake --build build
cmake --install build

set PATH=C:\Users\Quickemu\Documents\FTorch\.ftorch\Lib\site-packages;%PATH%
set PATH=C:\Program Files (x86)\FTorch\bin;%PATH%
set PATH=C:\Users\Quickemu\Documents\FTorch\.ftorch\Lib\site-packages\torch\lib;%PATH%

cd C:\Users\Quickemu\Documents\FTorch\examples\1_SimpleNet\build

simplenet_infer_fortran.exe ..\saved_simplenet_model_cpu.pt

The set PATH did not worked for me and I had the copy the requested dll's into the \examples\1_SimpleNet\build folder.

[TomMelt]

@xistocarlos thanks for the feedback. That's very helpful. I will aim to get this merged asap and I'll update the windows installation instructions so hopefully it gets easier for future windows users 🤞

Please stay in touch and let us know how you get on with FTorch or if you need any more help in future.

Can I ask what you are using it for? It's optional obviously but we love hearing how people are using our tool and it's nice to have examples of it being used in the community.

[xistocarlos]

I m the one that need to thank you guys for your support.

In our research we investigate novel propulsion systems for aircraft. We have our in-house simulation platform called GESTPAN, which is written in Fortran.

We are now expanding it to support the integration of surrogate models, trained with Pytorch. Our first attempt will be a neural network model trained on a new component 8 inputs and 8 outputs (2048 hidden dimensions and 0 layers). We are doing this as part of project MINIMAL (https://www.minimal-aviation.eu/).

I m now in the process of looking into your examples and expanded it for that purpose.

Is there any source we can cite from your work when describing the lib implementation?

[TomMelt]

That's great. Really good to hear it's being used to reduce CO2 emissions, especially considering I/we work for the Institute of Computing for Climate Science 🙌 🌍

We have a JOSS paper that is close to submission (@jatkinson1000 can possibly update once we have a DOI).

Please stay in touch, and if you need any help getting it working feel free to raise another issue.