experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Add TODO for defining parallel_execute op via tablegen.

Browse Source 
PiperOrigin-RevId: 292601015
Change-Id: I34775078ad88a77e0303ed1547bb147db3cedc12
This commit is contained in:
A. Unique TensorFlower 2020-01-31 13:03:46 -08:00 committed by TensorFlower Gardener
parent 688c4c2842
commit 67ab48e1f9
1 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
tensorflow/compiler/mlir/tensorflow/ir/tf_device.h
View File

@ -42,6 +42,9 @@ class TensorFlowDeviceDialect : public Dialect {
#define GET_OP_CLASSES #define GET_OP_CLASSES
#include "tensorflow/compiler/mlir/tensorflow/ir/tf_device.h.inc" #include "tensorflow/compiler/mlir/tensorflow/ir/tf_device.h.inc"
// TODO(b/148642767): Use tablegen to define tf_device.parallel_execute op once
// variadic regions can be expressed in tablegen.
//
// ParallelExecute op concurrently executes variadic number of regions. Regions // ParallelExecute op concurrently executes variadic number of regions. Regions
// must represent separate sets of instructions to execute concurrently. In // must represent separate sets of instructions to execute concurrently. In
// order to represent concurrently executed regions with dependencies, multiple // order to represent concurrently executed regions with dependencies, multiple