diff options
| author |  Bixia Zheng <bixia@google.com> | 2018-02-12 16:56:28 -0800 |
|---|---|---|
| committer |  TensorFlower Gardener <gardener@tensorflow.org> | 2018-02-12 16:59:58 -0800 |
| commit | 929e3ee91ecf7f9685b50fa1681f39d9b25e568b (patch) | |
| tree | f504d222b6255c9d48d67341282d27374ee26c5f /tensorflow/core/example | |
| parent | 96564330fb0508a50a0515be11c9202c64b0f5b7 (diff) | |
[XLA:GPU] Extend the CustomCall for cudnn convolutions to represent
tensor_ops_enabled.
The convolution algorithms returned from the stream executor have a flag
for whether tensor_ops is enabled. This flag is used when running each
algorithm during auto-tunning. However, this flag is not currently represented
in the CustomCall representing the auto-tune result. As a result, the algorithm
may be run differently after auto-tune.
This change adds a constant to the CustomCall for cudnn convolution algorithm
selected by auto-tune, to represent whether tensor_ops is enabled during
auto-tune. This information is used by convolution thunk to ensure that the
algorithm is run with the same flag after auto-tune.
PiperOrigin-RevId: 185458497
Diffstat (limited to 'tensorflow/core/example')
0 files changed, 0 insertions, 0 deletions