torch-cluster installation fails on Windows 10 with Python 3.7, PyTorch 1.2, CUDA 10.0

[chadykamar]

Environment looks like so

$ pip list
decorator             4.4.0
googledrivedownloader 0.4
h5py                  2.9.0
isodate               0.6.0
joblib                0.13.2
networkx              2.3
numpy                 1.17.1
pandas                0.25.1
Pillow                6.1.0
pip                   19.2.3
plyfile               0.7
pyparsing             2.4.2
python-dateutil       2.8.0
pytz                  2019.2
rdflib                4.2.2
scikit-learn          0.21.3
scipy                 1.3.1
setuptools            41.2.0
six                   1.12.0
torch                 1.2.0
torch-geometric       1.3.1
torch-sparse          0.4.0
torchvision           0.4.0
wheel                 0.33.6
$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Sat_Aug_25_21:08:04_Central_Daylight_Time_2018
Cuda compilation tools, release 10.0, V10.0.130
$ python --version
Python 3.7.4
$ python -c "import torch; print(torch.version.cuda); print(torch.cuda.is_available()); print(torch.__version__)"
10.0
True
1.2.0
$ env | grep CUDA
CPATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include
CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
CUDA_PATH_V10_0=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
NVCUDASAMPLES10_0_ROOT=C:\ProgramData\NVIDIA Corporation\CUDA Samples\v10.0
NVCUDASAMPLES_ROOT=C:\ProgramData\NVIDIA Corporation\CUDA Samples\v10.0
PATH=/c/Users/ckamar/Developer/SimGNN/venv/Scripts:/c/tools/Cmder/vendor/conemu-maximus5/ConEmu/Scripts:/c/tools/Cmder/vendor/conemu-maximus5:/c/tools/Cmder/vendor/conemu-maximus5/ConEmu:/c/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.0/bin:/c/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v10.0/libnvvp:...

It prints out a lot of warnings and ends up failing on this:

 graclus_kernel.cu
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1379): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1437): note: see reference to class template instantiation 'ska::flat_hash_map<K,V,H,E,A>' being compiled
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1383): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1391): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1473): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1510): note: see reference to class template instantiation 'ska::flat_hash_set<T,H,E,A>' being compiled
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1478): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1482): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1486): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    c:/users/ckamar/developer/simgnn/venv/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1490): error C3203: 'templated_iterator': unspecialized class template can't be used as a template argument for template parameter '_Ty1', expected a real type
    error: command 'C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v10.0\\bin\\nvcc.exe' failed with exit status 2
    ----------------------------------------
ERROR: Command errored out with exit status 1: 'c:\users\ckamar\developer\simgnn\venv\scripts\python.exe' -u -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'C:\\Users\\ckamar\\AppData\\Local\\Temp\\pip-install-6yaa4l2t\\torch-cluster\\setup.py'"'"'; __file__='"'"'C:\\Users\\ckamar\\AppData\\Local\\Temp\\pip-install-6yaa4l2t\\torch-cluster\\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' install --record 'C:\Users\ckamar\AppData\Local\Temp\pip-record-9schp724\install-record.txt' --single-version-externally-managed --compile --install-headers 'c:\users\ckamar\developer\simgnn\venv\include\site\python3.7\torch-cluster'

Any help is appreciated. Thanks.

[rusty1s]

Are you able to compile the CUDA samples with nvcc? Can you test to install the official extension example of PyTorch?

[chadykamar]

Sorry for not answering sooner. That was indeed the issue. I never quite managed to get nvcc working on Windows, but both it and the whole pytorch_geometric stack works like a charm on Linux :)