the regression is caused when only enable CL_UNORM_INT8 for CL_RG,
the reason is that during the image copy implementation with internal
kernel, all formats are considerd as integer format, it becomes
unknown since CL_UNSIGNED_INT8 is not enabled yet.
Signed-off-by: Guo Yejun
---
src/cl_image.c
The half float can work for BSW,
I will send a standalone patch to enable it later.
On 2015年06月19日 15:18, Yang, Rong R wrote:
One concern: Does cherryview support half?
The other part of the patchset LGTM.
-Original Message-
From: Beignet [mailto:beignet-boun...@lists.freedesktop.org
On 07/01/2015 07:28 AM, He, Junyan wrote:
So far as we know, beignet is thread safe.
Every thead has its own command buffer and do not have
relationship with each others.
Do you use subbuffer to divide the image?
No, as far as I can tell I don't use anything special or advanced.
If you can prov
Selecting a reasonable threshold seems a difficult job. I am ok with 1024.
LGTM.
> -Original Message-
> From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On Behalf Of
> Yang Rong
> Sent: Wednesday, July 01, 2015 11:31 AM
> To: beignet@lists.freedesktop.org
> Cc: Yang, Rong R
> S
This is special optimization for below situation:
bb1:
...
bb2:
x = phi [x1, bb1], [x2, bb2]
x2 = x+1;
after de-ssa:
bb2:
mov x, x-copy
add x2, x, 1
mov x-copy, x2
obviously x2, x-copy and x2 can be mapped to same virtual register.
Signed-off-by: Ruiling Song
---
backend
