Permalink
Please
sign in to comment.
Browse files
Revert "Add dynamic buffer support to OCL Backend (#3765)" (#3784)
Summary: This reverts commit bd69664. I had thought that I had gotten the last POCL issue in #3765, but I had not. Reverting to fix the OCL build. Honestly this last issue (AMD/POCL requires sub buffers to aligned) seems to torpedo the whole idea, I can't think of any way to handle Glow TensorViews on the host - which means passing buffer + offset everywhere we pass a buffer below. Essentially this would mean rewriting the whole thing. Very frustrating since that alignment restriction on subBuffers makes no sense, and no other OCL implementation has it. Pull Request resolved: #3784 Differential Revision: D18480248 Pulled By: nickgg fbshipit-source-id: 9b05009ea901a0f477805e6c946faac34d9bc303
- Loading branch information
Showing
with
490 additions
and 232 deletions.
- +1 −1 .circleci/build.sh
- +93 −127 lib/Backends/OpenCL/OpenCL.cpp
- +5 −13 lib/Backends/OpenCL/OpenCL.h
- +2 −34 lib/Backends/OpenCL/OpenCLDeviceManager.cpp
- +0 −11 lib/Backends/OpenCL/OpenCLDeviceManager.h
- +373 −39 lib/Backends/OpenCL/kernels.cl
- +7 −2 lib/Backends/OpenCL/kernels_fwd_conv.cl
- +9 −5 lib/Backends/OpenCL/kernels_fwd_quantized_conv.cl

Oops, something went wrong.
0 comments on commit
5dac6b2