[ROCm] Interface changes for StreamExecutor to support both CUDA and ROCm

1) StreamInterface::CudaStreamMemberHack()

Despite the fact that StreamExecutor and GPU common runtime are largely
orthogonal, this particular routine in StreamExecutor is used in GPU common
runtime and a couple of other operators. In this commit it's renamed as
StreamInterface::GpuStreamMemberHack() and their call sites are also changed.

2) StreamExecutorInterface::CudaContextHack()

This member is renamed to StramExecutorInterface::GpuContextHack().

Changes introduced in this commit includes:

- some StreamExecutor interfaces and CUDA implementation
- GPU common runtime related to interface changes in StreamExecutor
- operators affected by interface changes in StreamExecutor

1 files changed, 1 insertions, 1 deletions

diff --git a/tensorflow/contrib/nccl/kernels/nccl_manager.cc b/tensorflow/contrib/nccl/kernels/nccl_manager.cc
index b1cb89391c..99fecf9651 100644
--- a/tensorflow/contrib/nccl/kernels/nccl_manager.cc
+++ b/tensorflow/contrib/nccl/kernels/nccl_manager.cc
@@ -445,7 +445,7 @@ void NcclManager::LoopKernelLaunches(NcclStream* nccl_stream) {
   se::Stream* comm_stream = nccl_stream->stream.get();
   ScopedActivateExecutorContext scoped_context(nccl_stream->executor);
   const cudaStream_t* cu_stream = reinterpret_cast<const cudaStream_t*>(
-      comm_stream->implementation()->CudaStreamMemberHack());
+      comm_stream->implementation()->GpuStreamMemberHack());
 
   while (true) {
     // Find collective to run.