[Beignet] [PATCH 4/5] GBE: Add datalayout and triple to ll files.

It can fix datalayout mismatch warning in llvm3.7. Signed-off-by: Yang Rong --- backend/src/libocl/src/ocl_barrier.ll | 3 +++ backend/src/libocl/src/ocl_clz.ll | 3 +++ 2 files changed, 6 insertions(+) diff --git a/backend/src/libocl/src/ocl_barrier.ll

[Beignet] [PATCH 2/5] GBE: Add llvm3.7 support.

Move all llvm relative includes to llvm_includes.hpp. Signed-off-by: Yang Rong --- backend/src/backend/gen_program.cpp | 4 + backend/src/llvm/ExpandConstantExpr.cpp | 7 +- backend/src/llvm/ExpandLargeIntegers.cpp | 21 +---

[Beignet] [PATCH 3/5] GBE: Move createStripAttributesPass before createInstructionCombiningPass.

Otherwise, createInstructionCombiningPass will convert some call to illegal instruction in llvm3.7, for example utest compiler_time_stamp and test_load_program_from_spir. Signed-off-by: Yang Rong --- backend/src/llvm/llvm_to_gen.cpp | 2 +- 1 file changed, 1

[Beignet] [PATCH] SKL: explicit disable media sampler DOP clock gate in PIPELINE_SELECT.

This bit maybe enable by sbios, explicit disable to save power. Signed-off-by: Yang Rong --- src/intel/intel_defines.h | 10 ++ src/intel/intel_gpgpu.c | 3 ++- 2 files changed, 8 insertions(+), 5 deletions(-) diff --git a/src/intel/intel_defines.h

[Beignet] [PATCH] Don't use cl_buffer_get_subdata in clEnqueueReadBuffer.

cl_buffer_get_subdata sometime is very very very slow in linux kernel, in skl and chv, and it is random. So temporary disable it, use map/copy/unmap to read. Should re-enable it after find root cause. Signed-off-by: Yang Rong --- src/cl_enqueue.c | 5 - 1 file

Re: [Beignet] [PATCH 1/2] GBE: continue to refine interfering check.

Ping for review. Thanks. On Sun, Sep 06, 2015 at 03:05:00PM +0800, Zhigang Gong wrote: > More aggresive interfering check, even if both registers are in > Livein set or Liveout set, they are still possible not interfering > to each other. > > Signed-off-by: Zhigang Gong

Re: [Beignet] [PATCH 1/5] GBE: refine Phi copy interfering check.

Ping for review. Thanks. On Tue, Sep 01, 2015 at 12:04:59PM +0800, Zhigang Gong wrote: > If the PHI source register's definition instruction uses the > phi register, it is not a interfere. For an example: > > MOV %phi, %phicopy > ... > ADD %phiSrcDef, %phi, tmp > ... > MOV %phicopy, %phiSrcDef >