Which platform does your code base on?
I've run ./utest_run compiler_workgroup_reduce_add_float on IVB and BDW, it
succeed on BDW but fail
on IVB.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Grigore Lupescu
Sent: Tuesday, March 1, 2016 5:2
/beignet_new/beignet/utests/compiler_workgroup_reduce.cpp, function
compiler_workgroup_reduce_add_float, line 239
-Original Message-
From: Lupescu, Grigore
Sent: Friday, March 4, 2016 4:51 PM
To: Weng, Chuanbo ; beignet@lists.freedesktop.org
Subject: RE: [Beignet] [PATCH 2/2] Workgroup reduce
LGTM, thanks.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
yan.w...@linux.intel.com
Sent: Monday, March 28, 2016 3:35 PM
To: beignet@lists.freedesktop.org
Cc: Yan Wang
Subject: [Beignet] [PATCH] Add condition checking of residuals because i
Please ignore this patch, I'll send out a new one.
-Original Message-
From: Weng, Chuanbo
Sent: Wednesday, August 31, 2016 12:34 AM
To: beignet@lists.freedesktop.org
Cc: Weng, Chuanbo
Subject: [PATCH] CL: Re-enable cl_khr_gl_sharing with existing egl extension.
In order to quer
l-platforms="x11,drm"
make
make install
-Original Message-
From: Yang, Rong R
Sent: Thursday, September 1, 2016 2:04 PM
To: Weng, Chuanbo ; beignet@lists.freedesktop.org
Cc: Weng, Chuanbo
Subject: RE: [Beignet] [PATCH] Runtime: re-enable cl_khr_gl_sharing with
existing egl ex
Seems can't send out email yesterday...
Re-send.
-Original Message-
From: Weng, Chuanbo
Sent: Monday, September 5, 2016 3:35 PM
To: Yang, Rong R ; beignet@lists.freedesktop.org
Subject: RE: [Beignet] [PATCH] Runtime: re-enable cl_khr_gl_sharing with
existing egl extension.
In ord
Hi Seva,
What Rong means is that Beignet will try to open drm device exposed in
/dev/dri/ (Actually, In non-X environment,
Beignet will try to open render node such as /dev/dri/renderD128, if failed it
will then try to open prime primary node such
as /dev/dri/card0).
Beignet can
est,
Seva
23.09.2016, 12:33, "Weng, Chuanbo" :
> Hi Seva,
> What Rong means is that Beignet will try to open drm device
> exposed in /dev/dri/ (Actually, In non-X environment, Beignet will try
> to open render node such as /dev/dri/renderD128, if failed it will the
Hi Rebecca,
Please see my comments below.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Rebecca N. Palmer
Sent: Saturday, October 8, 2016 9:27 PM
To: beignet@lists.freedesktop.org
Subject: [Beignet] [PATCH] Utests: Allow testing cl_in
Hi Rebecca,
This version LGTM except some points need to be minor refined. Just see
my comments below.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Rebecca N. Palmer
Sent: Wednesday, October 12, 2016 5:50 AM
To: Weng, Chuanbo
ush Rebecca's patch. Otherwise there
will be FAIL on non-IVB platform.
Thanks,
Chuanbo Weng
-Original Message-
From: Rebecca N. Palmer [mailto:rebecca_pal...@zoho.com]
Sent: Thursday, October 13, 2016 6:38 AM
To: Weng, Chuanbo ; beignet@lists.freedesktop.org
Subject: Re: [Be
Hi Rebecca,
cl_khr_gl_sharing is partially supported (the most commonly used part).
I have just sent out a patchset that show how to use
cl_khr_gl_sharing on Beignet. The patchset also detect GL and EGL version and
turn cl_khr_gl_sharing ON if version >= 13.0.0.
So the first hunk
This patch LGTM, and I did a basic test on this patch. Thanks for your patch.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Rebecca N. Palmer
Sent: Monday, January 23, 2017 7:11 AM
To: beignet@lists.freedesktop.org
Subject: [Beignet] [PATCH]
Hi Sateesh,
Yes, vme now is just supported on 3rd generation.
And we are implementing vme on 6th generation (SKYLAKE) platform
firstly, because it’s a common used platform currently.
4th generation is also a little old.
Thanks,
Chuanbo Weng
From: Beignet [mailto:bei
:17 PM
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org
Subject: Re: [Beignet] VME support for Intel 4th generation (Haswell)
Hi Chuanbo
Understood and it's good news that this was planned for future (since I haven't
encountered any posts on motion estimation in Beignet).
I n
Hi
There is a known bug in beignet. I've attach the reproduce case.
Just run:
gcc bool_bug.cpp -lcl -o bool_bug
./bool_bug
And the result is as below:
ASSERTION FAILED: TODO Boolean values cannot escape their definition
basic block
Thanks,
Chuanbo Weng
bool_
@lists.freedesktop.org
Cc: Zou, Nanhai; Gong, Zhigang; Weng, Chuanbo
Subject: [Beignet] [PATCH 3/3] Add a test case that trigger vstore3 bug.
When there is no workgroup, the vstore3 will store incorrect value to buffer.
Signed-off-by: Chuanbo Weng
---
kernels/compiler_vstore3_no_workgroup.cl | 11
Sent: Thursday, August 29, 2013 2:42 PM
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org; Zou, Nanhai; Gong, Zhigang
Subject: Re: [Beignet] [PATCH] Add a test case that trigger vstore3 bug.
Hi chuanbo,
I just tried this case on my machine. It just passed as below:
compiler_vstore3_no_work
Hi Nikita,
As the mail that you found said, currently libva/OpenCL buffer sharing
is integrated
in libva. Please note, you should clone the correct libva repo as following
(not official libva):
git://people.freedesktop.org/~yakuiz/libva
Then you will see cl_branch in the tree. Then you ca
ver's official
staging branch and apply this patch to libva-intel-driver:
http://lists.freedesktop.org/archives/libva/2014-June/002463.html
Thanks,
Chuanbo Weng
-Original Message-
From: Nikita Orlov [mailto:nikitos1...@yandex.ru]
Sent: Tuesday, June 17, 2014 17:44
To: Weng, Chuanbo; be
ramastu [mailto:murama...@gmail.com]
Sent: Friday, August 15, 2014 15:54
To: Weng, Chuanbo
Subject: About "beignet-openCL and libva/VAAPI"
Hi! Chuanbo Weng,
I'm find a way to share memory between libva and opencl without memory copy
and just read the post you wrote
.
Also cc Yakui.
Thanks,
Chuanbo Weng
From: Tu Muramastu [mailto:murama...@gmail.com]
Sent: Thursday, August 21, 2014 18:00
To: Weng, Chuanbo
Subject: Re: About "beignet-openCL and libva/VAAPI"
Hi~ Chuanbo Wen,
I've tried to fill the attribute as your example and th
Hi,
What you need is the way to connect libva decode and ocl
post-processing by
gpu buffer sharing. You just need to:
1. Decode your video/picture to a VASurface.
2. Use vaAcquireBufferHandle to get VASurface’s handle.
3. Use clCreateBufferFromLibvaIntel/clCrea
Please ignore this patchset. I'll send out a new version.
-Original Message-
From: Weng, Chuanbo
Sent: Tuesday, December 16, 2014 20:17
To: beignet@lists.freedesktop.org
Cc: Zou, Nanhai; Zhao, Yakui; Li, Jocelyn; Weng, Chuanbo
Subject: [PATCH 2/2] Add document to describe the detia
Hi all,
Due to the test image size issue in the patch, I have made a
new version of patches
in order to switch to a smaller test image. If the patch is directly sent to
mailing-list with git
send-email, the test image can't be encoded in the correct format. So I put the
patch on
Ping for review.
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of Weng,
Chuanbo
Sent: Monday, December 22, 2014 18:38
To: beignet@lists.freedesktop.org
Cc: Zou, Nanhai; Gong, Zhigang
Subject: [Beignet] version 2 of libva buffer sharing patch set
Hi all
hould
download libva source code by themselves.
-Original Message-
From: Zhigang Gong [mailto:zhigang.g...@linux.intel.com]
Sent: Wednesday, January 07, 2015 8:49
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org; Zou, Nanhai; Gong, Zhigang
Subject: Re: [Beignet] version 2 of libva b
That's ok. I'll re-org the files and import libva as submodule.
Any other comments for this patchset?
-Original Message-
From: Zhigang Gong [mailto:zhigang.g...@linux.intel.com]
Sent: Thursday, January 08, 2015 10:25
To: Weng, Chuanbo
Cc: Gong, Zhigang; beignet@lists.freed
Hi all,
Please fetch the 3rd version of patchset from github using the following
command:
git clone https://github.com/Chuanbo-Weng/libva-buffer-sharing-patches.git
Or visit:
https://github.com/Chuanbo-Weng/libva-buffer-sharing-patches
*For reviewers:
Because we add libva as submodule
Hi Yejun,
Could you please have a try and review this patchset? Because
you have some
experience of libva buffer sharing.
Thanks,
Chuanbo Weng
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of Weng,
Chuanbo
Sent: Wednesday, January 21, 2015 11:07
To
Sorry, forgot to add new patch files to repo.
Now it's updated. Please try again.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Zhigang Gong
Sent: Friday, January 23, 2015 12:54
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org
Subjec
Hi all,
Please fetch the 4th version of patchset from github using the following
command:
git clone https://github.com/Chuanbo-Weng/libva-buffer-sharing-patches.git
Or visit:
https://github.com/Chuanbo-Weng/libva-buffer-sharing-patches
*For reviewers:
Because we add libva as submodule
, Yejun
Sent: Monday, January 26, 2015 13:58
To: Weng, Chuanbo; beignet@lists.freedesktop.org
Subject: RE: version 4 of libva buffer sharing patchset
Some comments:
- The examples build is bound closely with libva, no libva no build
examples. Any expectation that other examples without libva
Ping for review.
-Original Message-
From: Weng, Chuanbo
Sent: Tuesday, January 27, 2015 14:19
To: beignet@lists.freedesktop.org
Cc: Weng, Chuanbo
Subject: [PATCH 1/2] Refine benchmark output.
Change time unit from int to double, becasue it's not precise enough for some
small wor
/s' and 'GB/s' are ok for me. But since your internal output
in vload_bench are
'GB/s', I think using 'GB/s' (with return double value ) are better.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Zhigang G
Removed and updated the patchset.
From: Guo, Yejun
Sent: Monday, February 02, 2015 14:11
To: Weng, Chuanbo; beignet@lists.freedesktop.org
Subject: RE: version 4 of libva buffer sharing patchset
Looks fine except one comment: OUTPUT_NV12_DEFAULT is not really used, we can
just remove it.
From
Thanks for your comment, I'll send out a new version to fix this error.
-Original Message-
From: Zhigang Gong [mailto:zhigang.g...@linux.intel.com]
Sent: Friday, February 13, 2015 10:18
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org
Subject: Re: [Beignet] [PATCH] Optimizati
This patch could fix the API/ get_kernel_arg_info fail case in Khronos OpenCL
1.2 conformance test.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
Zhigang Gong
Sent: Wednesday, March 18, 2015 15:08
To: beignet@lists.freedesktop.org
Cc: Gong,
esday, April 01, 2015 8:52
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org
Subject: RE: [PATCH 1/3] add benckmark for copy data from buffer to image.
Hi Chuanbo,
Please review this patchset for me.
It adds aligned copy data from buffer to image 2d (only image 2d to buffer
before).
Luo Xionghu
See my comments below for cl_internal_copy_buffer_to_image_2d_align16.cl.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
xionghu@intel.com
Sent: Wednesday, April 01, 2015 8:48
To: beignet@lists.freedesktop.org
Cc: Luo, Xionghu
Subject: [Be
One warning when running "git am" command:
warning: 1 line adds whitespace errors.
And the words of the patch description "thus 16 bytes can be read by one work
item " should be changed
to "thus 16 bytes can be written by one work item ".
I think Zhigang can help to do this minor modification befo
LGTM.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
xionghu@intel.com
Sent: Wednesday, April 01, 2015 13:11
To: beignet@lists.freedesktop.org
Cc: Luo, Xionghu
Subject: [Beignet] [patch v2 2/2] add benckmark for copy data from buffer to
i
al Message-
From: Luo, Xionghu
Sent: Thursday, April 02, 2015 11:16
To: Weng, Chuanbo; beignet@lists.freedesktop.org
Subject: RE: [Beignet] [patch v2 1/2] Optimization of
clEnqueueCopyBufferToImage for 16 aligned case.
For CopyBufferToImage kernel, "thus 16 bytes can be read by one
9:45
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org; Yuan, Feng
Subject: Re: [Beignet] [PATCH v2 3/3] Add document to describe the detials of
v4l2 buffer sharing.
One major comment is we should not introduce such an extension without
considering the queue. The correct APIs should be some
ok. As we discussed, you'd better send out a new version:)
-Original Message-
From: Luo, Xionghu
Sent: Friday, April 03, 2015 11:17
To: Weng, Chuanbo; beignet@lists.freedesktop.org
Subject: RE: [Beignet] [patch v2 1/2] Optimization of
clEnqueueCopyBufferToImage for 16 aligned cas
Hi Zhenyu,
Please my comments below.
-Original Message-
From: Zhenyu Wang [mailto:zhen...@linux.intel.com]
Sent: Thursday, April 09, 2015 12:31
To: Weng, Chuanbo
Cc: beignet@lists.freedesktop.org; Yuan, Feng
Subject: Re: [Beignet] [PATCH v3 1/2] Add example to show v4l2 buffer
This patchset looks good to me.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
xionghu@intel.com
Sent: Thursday, April 09, 2015 11:38
To: beignet@lists.freedesktop.org
Cc: Luo, Xionghu
Subject: [Beignet] [patch v3 1/2] Optimization of clEn
Please ignore this patchset, I will send out a new version soon.
-Original Message-
From: Weng, Chuanbo
Sent: Monday, September 07, 2015 13:01
To: beignet@lists.freedesktop.org
Cc: Weng, Chuanbo
Subject: [PATCH v2 1/4] Add built-in function __gen_ocl_vme.
__gen_ocl_vme is used for
Agreed. Good suggestion.
I will refine the patch.
-Original Message-
From: Guo, Yejun
Sent: Friday, September 18, 2015 11:47
To: Zhao, Yakui; Weng, Chuanbo
Cc: beignet@lists.freedesktop.org
Subject: RE: [Beignet] [PATCH] Add extension clCreateMemObjectFromFdIntel to
create cl memory
Ping for review, thanks.
-Original Message-
From: Weng, Chuanbo
Sent: Monday, September 21, 2015 16:20
To: beignet@lists.freedesktop.org
Cc: Wu, Zhiwen; Weng, Chuanbo
Subject: [PATCH v2 1/2] Add extension clCreateBufferFromFdINTEL to create cl
buffer by external buffer object'
Hi Ruiling,
As we discussed before, I have refined code to handle both simd8 and
simd16 in backend and ocl kernel. Please
confirm if no problem so that this patchset can be pushed. Thanks!
-Original Message-
From: Weng, Chuanbo
Sent: Friday, November 06, 2015 11:28
To: beignet
Ping for review, thanks!
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of Weng,
Chuanbo
Sent: Tuesday, October 27, 2015 10:16
To: beignet@lists.freedesktop.org
Cc: Wu, Zhiwen
Subject: Re: [Beignet] [PATCH v2 1/2] Add extension
52 matches
Mail list logo
