Nathan Luehr
2642e93e6c
Use correct cudart soname in GetDsoHandle
2020-10-01 19:58:42 -05:00
Nathan Luehr
28feb4df0d
Enable build with CUDA 11
2020-05-14 16:03:21 -05:00
Christian Sigg
1588f45ee5
Remove no longer used compute capability list macro in C++ header.
...
PiperOrigin-RevId: 308005305
Change-Id: I71f1f0ab3b1fd33958bf28c563b0fdc3c68db5a4
2020-04-23 02:12:52 -07:00
A. Unique TensorFlower
8e9ca6617b
Change how ./configure and bazel finds CUDA paths.
...
Use a single python script (third_party/gpus/find_cuda_config.py) from configure.py and the different *_configure.bzl scripts to find the different CUDA library and header paths based on a set of environment variables.
PiperOrigin-RevId: 243669844
2019-04-15 12:55:59 -07:00
Jingyue Wu
c1238db44e
CudaRoot() returns the configured CUDA toolkit path.
...
Fixes the problem of XLA being unable to find libdevice files if not executed
through bazel. For example,
http://stackoverflow.com/questions/41729019/notfounderror-running-tensorflow-xla-example-libdevice-compute-35-10-bc/41800414#41800414
Change: 146329331
2017-02-01 22:28:06 -08:00
A. Unique TensorFlower
2c598e874e
Merge changes from github.
...
Change: 131310818
2016-08-25 11:32:33 -07:00
