search

openmm / NNPOps

High-performance operations for neural network potentials
Other
81 stars 17 forks source link

[DO NOT MERGE] Test if we can build with pytorch 1.11 #57

Closed mikemhenry closed 2 years ago

mikemhenry commented 2 years ago

Running into this error when building with pytorch 1.11 on cuda 10, going to see if first we can build with 1.11

mikemhenry commented 2 years ago

@raimis

The combination of GCC 7.5.0 + CUDA 10.2 + pytorch 1.11 gives me this error: 

CMake Error in CMakeLists.txt:
  Target "NNPOpsPyTorch" requires the language dialect "CUDA17" (with
  compiler extensions).  But the current compiler "NVIDIA" does not support
  this, or CMake does not know the flags to enable it.

What confuses me is pytorch 1.10 builds just fine, but I don't see anything in the pytorch 1.11 release notes that make it seem like this would be a problem.

https://github.com/conda-forge/nnpops-feedstock/pull/3