Fixed the crash for --cu-lossless command. The crash was due to divide by
zero issue. Successfully ran smoke test on it.
On Mon, May 30, 2016 at 1:54 PM, wrote:
> # HG changeset patch
> # User Ashok Kumar Mishra
> # Date 1464596647 -19800
# HG changeset patch
# User Ashok Kumar Mishra
# Date 1464596647 -19800
# Mon May 30 13:54:07 2016 +0530
# Node ID 21c9f83bcdac2d2c5beca9c7349e397460dd94d5
# Parent aeade2e8d8688ebffb8455b8948d89d6a72e2c38
[OUTPUT CHANGED]SAO: convert sao rdo cost calculation from
This patch crashes cu-lossless commandlines.
On Fri, May 27, 2016 at 1:21 PM, wrote:
> # HG changeset patch
> # User Ashok Kumar Mishra
> # Date 1464335437 -19800
> # Fri May 27 13:20:37 2016 +0530
> # Node ID
# HG changeset patch
# User Ashok Kumar Mishra
# Date 1464335437 -19800
# Fri May 27 13:20:37 2016 +0530
# Node ID a98d2f032a1cfa4c5c279fe48bdf4b96b4907b45
# Parent 2a379fa9b5e85fa7c5240d231edb2396e2c4eb44
[OUTPUT CHANGED]SAO: convert sao rdo cost calculation from
The 10bpp build fails (lambda2) - that means the output effect on 10-bit
input has not been checked?
On Thu, May 26, 2016 at 6:05 PM, Ashok Kumar Mishra <
as...@multicorewareinc.com> wrote:
> The reason for output change is quite obvious. lambda value, cost
> calculation and comparison are
> on
The reason for output change is quite obvious. lambda value, cost
calculation and comparison are
on fixed point.
On Thu, May 26, 2016 at 5:54 PM, Ashok Kumar Mishra <
as...@multicorewareinc.com> wrote:
> this is a output changing patch. is there any other issue apart from
> changing output?
>
>
this is a output changing patch. is there any other issue apart from
changing output?
On Thu, May 26, 2016 at 5:46 PM, Deepthi Nandakumar <
deep...@multicorewareinc.com> wrote:
>
>
> On Wed, May 4, 2016 at 7:39 PM, wrote:
>
>> # HG changeset patch
>> # User Ashok
On Wed, May 4, 2016 at 7:39 PM, wrote:
> # HG changeset patch
> # User Ashok Kumar Mishra
> # Date 1462355258 -19800
> # Wed May 04 15:17:38 2016 +0530
> # Node ID 70a0888d0703a35b0c3c3a57f96931d0767eb470
> # Parent
Is there an impact of this on efficiency/performance? Do you have some
results that you can share?
Pradeep Ramachandran, PhD
Solution Architect at www.multicorewareinc.com/
Visiting Professor at www.cse.iitm.ac.in/
pradeeprama.info/
Ph: +91 99627 82018
On Wed, May 4, 2016 at 7:09 AM,
# HG changeset patch
# User Ashok Kumar Mishra
# Date 1462355258 -19800
# Wed May 04 15:17:38 2016 +0530
# Node ID 70a0888d0703a35b0c3c3a57f96931d0767eb470
# Parent 9f27620a948b67498056246b97db72bebac99218
[OUTPUT CHANGED]SAO: convert sao rdo cost calculation from
Results after converting from float to int.
*Before(crowd_run, medium)*
encoded 500 frames in 76.15s (6.57 fps), 9685.07 kb/s, Avg QP:37.80, Global
PSNR : 31.045, SSIM Mean Y: 0.8417967 ( 8.008 dB)
*After float to fixed*
encoded 500 frames in 75.17s (6.65 fps), 9689.38 kb/s, Avg QP:37.79, Global
# HG changeset patch
# User Ashok Kumar Mishra
# Date 1459928537 -19800
# Wed Apr 06 13:12:17 2016 +0530
# Node ID 511241d3ee7d6d53a999d46881e7921a601fa0e9
# Parent 33ff2e5f6eb7c8cf4f3edaa265762f32aa9b6f0f
[OUTPUT CHANGED]SAO: convert sao rdo cost calculation from
12 matches
Mail list logo