projects / platform / upstream / tvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 0d2b4bc) | patch
Make batch matrix multiplication on GPU tunable (#5752)
authorThomas Viehmann <tv.code@beamnet.de>
Thu, 11 Jun 2020 16:38:46 +0000 (18:38 +0200)
committerGitHub <noreply@github.com>
Thu, 11 Jun 2020 16:38:46 +0000 (09:38 -0700)
commit0abcad137448d6ef786fa90010d8320193c68f6f
treeae9d267151b54f1dece4f223e349363bf81b1bcdtree | snapshot
parent0d2b4bcd63ddc5d8e38c7be67180a0de6e41592ecommit | diff
Make batch matrix multiplication on GPU tunable (#5752)

This is primarily aimed at the AMD GPU backend and done as part
of a project for AMD, but should work for all users of the GPU
schedule.
python/tvm/relay/op/strategy/cuda.py diff | blob | history
topi/python/topi/cuda/batch_matmul.py diff | blob | history
topi/tests/python/test_topi_batch_matmul.py diff | blob | history
Domain: Machine Learning / ML Framework;