| d0ae1064ed0bb4bd1aed00afd4235f4dd5c853f0 |
|
01-Dec-2017 |
Max Galkin <maxgalkin@google.com> |
Prefix inaccurate costs with "~" in VirtualScheduler verbose log.
Fix some inaccurate estimates exposed by this approach:
- propagate the inaccuracy flag when merging device stats;
- estimate Const as no-op;
- estimate RandomUniform, Relu and Softmax as element-wise;
- consider estimates accurate for known element-wise ops in op_level_cost_estimator.
PiperOrigin-RevId: 177643976
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| 40dee372e3ee844c4746baa914c07b9c582a2ce7 |
|
27-Sep-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Define OpContext and use it for OpLevelCostEstimator.
This CL does not add any functionality (except GraphDef's function library pointer is passed to
OpContext), but we can later add additional fields to OpContext struct for extending
VirtualCluster, Scheduler, Placer, and others.
PiperOrigin-RevId: 170157235
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| 74057361032bc4d9a9fbfced1f433b06c06c09ec |
|
21-Jul-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add option to use compute_memory_overlap; if true, use max of memory_cost and compute_cost, instead of sum for op level cost in analytical cost estimator.
PiperOrigin-RevId: 162782658
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| 19b4ccd95fa667adf6240d24368022594f320e73 |
|
10-Jun-2017 |
Yao Zhang <yaozhang@google.com> |
Estimate cost for element wise ops and a minimum compute cost for dummy execution time.
PiperOrigin-RevId: 158585464
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| 74220616c3b6739b89c352f25fb6dccf483f7fab |
|
06-Jun-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Set device cores and frequency in op_level_cost_estimator_test,
to avoid asan error about assigning inf to int64 (this comes
in from a divide-by-0).
PiperOrigin-RevId: 158155488
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| a3ba225d5b327013709a1732688bfd4346b3c86e |
|
30-May-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add BatchMatMul execution cost prediction
PiperOrigin-RevId: 157487507
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| 87ba9f5370c0f7068760f9536979d9183f6dfe9c |
|
06-May-2017 |
Benoit Steiner <bsteiner@google.com> |
Return more complete device information from the GetDevices() method of grappler clusters.
Change: 155264843
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|
| 9b336b4a33158061535fd6ba4973605248055b69 |
|
04-May-2017 |
Benoit Steiner <bsteiner@google.com> |
Open sourced op level cost prediction
Change: 155123817
/external/tensorflow/tensorflow/core/grappler/costs/op_level_cost_estimator_test.cc
|