projects / platform / upstream / tensorflow.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2808c3f) | patch
Remove the inter-op thread pool
authorSanjoy Das <sanjoy@google.com>
Thu, 26 Apr 2018 22:33:38 +0000 (15:33 -0700)
committerTensorFlower Gardener <gardener@tensorflow.org>
Thu, 26 Apr 2018 22:36:18 +0000 (15:36 -0700)
commitab5de487813b4849dfb5415ee60595654dff06be
tree47ec64e0e02eea2e5fe0fae67e056f343dc905f7tree | snapshot
parent2808c3f05f7713ff1ab20f365e986a4651180376commit | diff
Remove the inter-op thread pool

Forgot about this in cl/194299356.  However, when I checked cl/194299356, I
found that we actually (incorrectly?) used the *intra* op thread pool in the
parallel CPU executable?  Does that mean the inter op thread pool was always
unused?

PiperOrigin-RevId: 194464734
tensorflow/compiler/xla/executable_run_options.cc diff | blob | history
tensorflow/compiler/xla/executable_run_options.h diff | blob | history
tensorflow/compiler/xla/python/local_computation_builder.cc diff | blob | history
tensorflow/compiler/xla/service/backend.cc diff | blob | history
tensorflow/compiler/xla/service/backend.h diff | blob | history
tensorflow/compiler/xla/service/hlo_runner.cc diff | blob | history
tensorflow/compiler/xla/service/service.cc diff | blob | history
tensorflow/compiler/xla/tests/local_client_test_base.cc diff | blob | history
Domain: Machine Learning / ML Framework;