Re: [Beignet] [PATCH 3/5] Remove intel_gpgpu_check_binded_buf_address()

2014-10-30 Thread Zhigang Gong
On Fri, Oct 31, 2014 at 02:19:54PM +0800, Zhenyu Wang wrote: > On 2014.10.27 17:53:51 +0800, Zhenyu Wang wrote: > > On 2014.10.27 12:57:59 +0800, Zhenyu Wang wrote: > > > On 2014.10.24 18:10:36 +0800, Zhigang Gong wrote: > > > > This assertion is just to make sure we will not get a NULL pointer > >

Re: [Beignet] [PATCH 3/5] Remove intel_gpgpu_check_binded_buf_address()

2014-10-30 Thread Zhenyu Wang
On 2014.10.27 17:53:51 +0800, Zhenyu Wang wrote: > On 2014.10.27 12:57:59 +0800, Zhenyu Wang wrote: > > On 2014.10.24 18:10:36 +0800, Zhigang Gong wrote: > > > This assertion is just to make sure we will not get a NULL pointer > > > for a normal buffer. The OpenCL spec doesn't give a very specific

Re: [Beignet] [PATCH] add llvm intrinsic call usub_with_overflow funtion.

2014-10-30 Thread Zhigang Gong
On Thu, Oct 30, 2014 at 10:59:37AM +0800, xionghu@intel.com wrote: > From: Luo Xionghu > > as llvm couldn't recognize the pattern of usub overflow, this usub with > is generated by calling the intrinsic function __builtin_usub_overflow; > also this type of uadd intrinsic funtion couldn't supp

Re: [Beignet] Copyright problems in test suite

2014-10-30 Thread Zhigang Gong
On Thu, Oct 30, 2014 at 11:18:47PM +, Rebecca N. Palmer wrote: > The Beignet test suite contains two groups of files that do not > appear consistent with the project's LGPL license: > > 1. Derived from the Len(n)a standard test image: > kernels/lenna128x128.bmp kernels/compiler_box_blur_float_

Re: [Beignet] beignet doesnt compile with LLVM 3.6

2014-10-30 Thread Yang, Rong R
HI, Paulo, Nice to meet you:) Do you use LLVM 3.6 to compile the Beignet’s C/C++ code? Sorry, I mistake that you only use LLVM 3.6 as Beignet’s OpenCL language compiler. I have quick look these compilation. For strict-aliasing warning, I think it is the clang’s issue. Because in beignet/backend

[Beignet] [PATCH 1/2] utest: change the box_blur_image to be identical to box_blur.

2014-10-30 Thread Zhigang Gong
Change box_blur_image to read integer type surface thus it could be totally identical to the box_blur thus they can share the same reference image. Signed-off-by: Zhigang Gong --- kernels/compiler_box_blur_image.cl | 8 +--- utests/compiler_box_blur_image.cpp | 2 +- 2 files changed, 6 inser

[Beignet] Copyright problems in test suite

2014-10-30 Thread Rebecca N. Palmer
The Beignet test suite contains two groups of files that do not appear consistent with the project's LGPL license: 1. Derived from the Len(n)a standard test image: kernels/lenna128x128.bmp kernels/compiler_box_blur_float_ref.bmp kernels/compiler_box_blur_ref.bmp This is not generally distribut

[Beignet] [PATCH] add utest function bswap.

2014-10-30 Thread xionghu . luo
From: Luo Xionghu Signed-off-by: Luo Xionghu --- kernels/compiler_bswap.cl |7 +++ utests/CMakeLists.txt |1 + utests/compiler_bswap.cpp | 44 3 files changed, 52 insertions(+) create mode 100644 kernels/compiler_bswap.cl create m

[Beignet] [PATCH] add llvm intrinsic call usub_with_overflow funtion.

2014-10-30 Thread xionghu . luo
From: Luo Xionghu as llvm couldn't recognize the pattern of usub overflow, this usub with is generated by calling the intrinsic function __builtin_usub_overflow; also this type of uadd intrinsic funtion couldn't support short/byte type overflow, we choose another way for the uadd kernel to genera

Re: [Beignet] beignet doesnt compile with LLVM 3.6

2014-10-30 Thread Paulo Dias
Hi Yang, nice to meet you. I agree with LLVM changing code from the development branch, but in this case, the compilation is failing because they switched to a stricter language convention, so it might be wiser to fix the code sooner then later. Im also creating beignet packages for future inclus