hlrs-software-stack/spack
7
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Use CudaPackage mixin for py-torch (#14540)

Browse source 
This PR adds CudaPackage in order to pick up the cuda/compiler conflicts
defined in CudaPackage.
This commit is contained in:
Glenn Johnson 2020-01-16 12:02:53 -06:00 committed by Adam J. Stewart
parent 8ad0be96aa
commit 5aeab7dbe5
1 changed files with 1 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
var/spack/repos/builtin/packages/py-torch/package.py
View file

@ -7,7 +7,7 @@
# TODO: try switching to CMakePackage for more control over build
class PyTorch(PythonPackage):
class PyTorch(PythonPackage, CudaPackage):
"""Tensors and Dynamic neural networks in Python
with strong GPU acceleration."""
@ -61,7 +61,6 @@ class PyTorch(PythonPackage):
version('0.4.0', tag='v0.4.0', submodules=True)
version('0.3.1', tag='v0.3.1', submodules=True)
variant('cuda', default=True, description='Enables CUDA build')
variant('cudnn', default=True, description='Enables the cuDNN build')
variant('magma', default=False, description='Enables the MAGMA build')
variant('fbgemm', default=False, description='Enables the FBGEMM build')