projects / platform / upstream / tvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 8bef595) | patch
Don't replace reduction init axis with new axis if bound to a thread. (#3408)
authorChristian Sarofeen <csarofeen@nvidia.com>
Mon, 12 Aug 2019 21:11:11 +0000 (17:11 -0400)
committerTianqi Chen <tqchen@users.noreply.github.com>
Mon, 12 Aug 2019 21:11:11 +0000 (14:11 -0700)
commitd482512d160566b618edf71159519aad502e8f62
treeaff5193a6622e88c65559fd30051b34a308f5160tree | snapshot
parent8bef595dededa2c6e34b43fa3e6d960da43a37d1commit | diff
Don't replace reduction init axis with new axis if bound to a thread. (#3408)

* Don't replace reduction init axis with new axis if bound to a thread.

* Linter.

* Reduce bind test case.

* Guard test on CUDA support.

* [CUDA TE TESTS] Add rfactor predicate test, add global bx and tx.

* [CUDA TE TESTS] Add loop partition test for simple rfactor case.
src/op/op_util.cc diff | blob | history
tests/python/unittest/test_codegen_cuda.py diff | blob | history
tests/python/unittest/test_pass_loop_partition.py diff | blob | history
Domain: Machine Learning / ML Framework;