[OpenCL] Implements BatchMatmul Op for SYCL#60

Open

lukeiwanski wants to merge 16 commits intobenoitsteiner:masterfrom

lukeiwanski:sycl_batchmatmull

Commits on Feb 25, 2017

Improvements to the SYCL device support

Luke Iwanski
authored and
benoitsteiner
committed
SYCL support improvements (#31)

Luke Iwanski
authored and
benoitsteiner
committed
Fixes & Version bump (#33)

Luke Iwanski
authored and
benoitsteiner
committed
Add problem descriptions and solutions (#35)

krikru
authored and
benoitsteiner
committed
Workaround for 0 bytes allocation for SYCL device (#42)

Luke Iwanski
authored and
benoitsteiner
committed
Sycl improvements (#44)

Luke Iwanski
authored and
benoitsteiner
committed
OpenCL Improvements

Luke Iwanski
authored and
benoitsteiner
committed

Commits on Feb 26, 2017

Commits on Mar 22, 2017

Minor fix to README.md (#50)

halfak
authored and
benoitsteiner
committed

Commits on Mar 27, 2017

Register SYCL implementations for random ops

ville-k
authored and
benoitsteiner
committed

Commits on Apr 4, 2017

OpenCL improvements

Luke Iwanski
authored and
benoitsteiner
committed
Avoid functions that might not be defined on SYCL device (#51)

ville-k
authored and
benoitsteiner
committed

Commits on Apr 5, 2017

[OpenCL] Fixes an issue caused by switch to aligned allocator for sycl buffer (#53)

Luke Iwanski
authored and
benoitsteiner
committed
[Build] Use gcc/g++ as a host compiler to avoid https://github.com/tensorflow/tensorflow/issues/8394 (#54)

Luke Iwanski
authored and
benoitsteiner
committed

Commits on Apr 6, 2017

[OpenCL] Implements BatchMatmul Op for SYCL
Luke
committed