[Beignet] [PATCH 1/2] handle simd8 and simd16 accrordingly for alu3.

From: Luo Xionghu for BDW+, there is no simd8 only limitation for three-source instructions. we just support mad three-source oprands now. Signed-off-by: Luo Xionghu --- backend/src/backend/gen8_encoder.cpp | 26 +- 1 file changed, 9 insertions(+), 17 deletions(-) diff

[Beignet] [PATCH 2/2] enable utest compiler_math_3op for mad test.

From: Luo Xionghu Signed-off-by: Luo Xionghu --- kernels/compiler_math_3op.cl | 2 +- utests/CMakeLists.txt| 1 + utests/compiler_math_3op.cpp | 9 + 3 files changed, 7 insertions(+), 5 deletions(-) diff --git a/kernels/compiler_math_3op.cl b/kernels/compiler_math_3op.cl index

Re: [Beignet] [PATCH 1/2] handle simd8 and simd16 accrordingly for alu3.

Hi Xionghu, I yesterday sent out a patch to handle the SIMD1 case for MAD, maybe you'd patch base on that, thanks. Yejun -Original Message- From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of xionghu@intel.com Sent: Wednesday, April 27, 2016 12:43 AM To: beig

Re: [Beignet] [PATCH v4 3/5] standalone utest for unified OpenCL implementation.

LGTM, pushed, thanks. It's better to add readme for it. > -Original Message- > From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of > xionghu@intel.com > Sent: Saturday, April 23, 2016 0:37 > To: beignet@lists.freedesktop.org > Cc: Luo, Xionghu > Subject: [Beigne

Re: [Beignet] [PATCH] runtime: handle PROGRAM_BUILD_GLOBAL_VARIABLE_TOTAL_SIZE

Pushed, thanks. > -Original Message- > From: Pan, Xiuli > Sent: Tuesday, April 26, 2016 14:43 > To: Song, Ruiling ; 'beignet@lists.freedesktop.org' > > Cc: Yang, Rong R > Subject: RE: [Beignet] [PATCH] runtime: handle > PROGRAM_BUILD_GLOBAL_VARIABLE_TOTAL_SIZE > > Ping to push! > > ---

Re: [Beignet] [PATCH] GBE: Fix destination grf register type for cmp instruction.

Pushed, thanks. > -Original Message- > From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of > Pan, Xiuli > Sent: Tuesday, April 26, 2016 14:40 > To: Song, Ruiling ; beignet@lists.freedesktop.org > Cc: Song, Ruiling > Subject: Re: [Beignet] [PATCH] GBE: Fix destination

[Beignet] [PATCH] GBE: try to avoid bank conflict in register allocator.

Signed-off-by: Ruiling Song --- backend/src/backend/gen_reg_allocation.cpp | 31 -- 1 file changed, 29 insertions(+), 2 deletions(-) diff --git a/backend/src/backend/gen_reg_allocation.cpp b/backend/src/backend/gen_reg_allocation.cpp index 89c53d4..ccb76a1 100644 ---

[Beignet] [PATCH] Backend: Chang scan limit for GVN pass

From: Pan Xiuli Set memdep-block-scan-limit into llvm context to avoid unfinished GVN pass. V2: Revert remove praser llvm first Signed-off-by: Pan Xiuli --- backend/src/backend/program.cpp | 22 +- 1 file changed, 21 insertions(+), 1 deletion(-) diff --git a/backend/src/ba

[Beignet] [PATCH] runtime: we can accept large program global variable.

We don't have explicit restriction in program global variable size. So, let's allow 128MB now, because 64KB is too small. Signed-off-by: Ruiling Song --- src/cl_gt_device.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/cl_gt_device.h b/src/cl_gt_device.h index 35a6ff5..