[x265] [PATCH] blockfill_s_16x16 avx2 asm code, performance improved 389.21 cycles -> 204.38 cycles

2014-09-29 Thread praveen
# HG changeset patch # User Praveen Tiwari # Date 1411980445 -19800 # Node ID 9a8552ea378500baa21b89b24d8aec99acf7cce2 # Parent 32f50df7fa7672f4c1818ddf3165b4bd243e0b10 blockfill_s_16x16 avx2 asm code, performance improved 389.21 cycles -> 204.38 cycles diff -r 32f50df7fa76 -r 9a8552ea3785 sourc

[x265] [PATCH] blockfill_s_32x32 avx2 asm code, performance improved 1354.05 cycles -> 705.81 cycles

2014-09-29 Thread praveen
# HG changeset patch # User Praveen Tiwari # Date 1411982102 -19800 # Node ID 543a79749978dc4aae2956788bb16e50d2ceca14 # Parent 9a8552ea378500baa21b89b24d8aec99acf7cce2 blockfill_s_32x32 avx2 asm code, performance improved 1354.05 cycles -> 705.81 cycles diff -r 9a8552ea3785 -r 543a79749978 sour

[x265] [PATCH 0 of 2 ] TComDataCU: replace with more CU structure details

2014-09-29 Thread santhoshini
___ x265-devel mailing list x265-devel@videolan.org https://mailman.videolan.org/listinfo/x265-devel

[x265] [PATCH 2 of 2] TComDataCU: replace getTotalNumPart() with CU structure details

2014-09-29 Thread santhoshini
# HG changeset patch # User Santhoshini Sekar # Date 1411983173 -19800 # Mon Sep 29 15:02:53 2014 +0530 # Node ID 594bad474d684cc23c84890f74b02ac42a85fc3e # Parent ed887d8ae5cd24b0c2317fb83b3c908be27e037a TComDataCU: replace getTotalNumPart() with CU structure details diff -r ed887d8ae5cd -

[x265] [PATCH 1 of 2] TComDataCU: replace getZorderIdxInCU() with encodeIdx of CU structure

2014-09-29 Thread santhoshini
# HG changeset patch # User Santhoshini Sekar # Date 1411972892 -19800 # Mon Sep 29 12:11:32 2014 +0530 # Node ID ed887d8ae5cd24b0c2317fb83b3c908be27e037a # Parent 32f50df7fa7672f4c1818ddf3165b4bd243e0b10 TComDataCU: replace getZorderIdxInCU() with encodeIdx of CU structure diff -r 32f50df7

Re: [x265] [PATCH 2 of 2] TComDataCU: replace getTotalNumPart() with CU structure details

2014-09-29 Thread Deepthi Nandakumar
On Mon, Sep 29, 2014 at 3:33 PM, wrote: > # HG changeset patch > # User Santhoshini Sekar > # Date 1411983173 -19800 > # Mon Sep 29 15:02:53 2014 +0530 > # Node ID 594bad474d684cc23c84890f74b02ac42a85fc3e > # Parent ed887d8ae5cd24b0c2317fb83b3c908be27e037a > TComDataCU: replace getTotalNum

Re: [x265] [PATCH 2 of 2] TComDataCU: replace getTotalNumPart() with CU structure details

2014-09-29 Thread Santhoshini Sekar
On Mon, Sep 29, 2014 at 3:41 PM, Deepthi Nandakumar < deep...@multicorewareinc.com> wrote: > > > On Mon, Sep 29, 2014 at 3:33 PM, wrote: > >> # HG changeset patch >> # User Santhoshini Sekar >> # Date 1411983173 -19800 >> # Mon Sep 29 15:02:53 2014 +0530 >> # Node ID 594bad474d684cc23c84890

Re: [x265] [PATCH 2 of 2] TComDataCU: replace getTotalNumPart() with CU structure details

2014-09-29 Thread Deepthi Nandakumar
Ok, thanks. cu->numPartitions = (NUM_CU_PARTITIONS >> ((g_maxLog2CUSize - cu->log2CUSize) * 2)); When a CU is initialised, it has NUM_CU_PARTITIONS partitions, (256 in the case of 64 size CTU). When a sub-CU is initialised or copied from the previous depth CU, it has 1/4th of the partitions of th

Re: [x265] [PATCH] Changes for loadCTUData

2014-09-29 Thread Deepthi Nandakumar
Ashok/Santhoshini - pls review. Does removing offsets affect any planned optimizations? On Sat, Sep 27, 2014 at 7:03 AM, wrote: > # HG changeset patch > # User David T Yuen > # Date 1411781537 25200 > # Node ID 85098db291ae133981419868685358227b8b1437 > # Parent 4b18a27b52ac69a16805c2b455d4f89

Re: [x265] [PATCH] Changes for loadCTUData

2014-09-29 Thread Santhoshini Sekar
As of now offset can be removed from CU structure. We don't have any planned optimization with it. If needed we can add it later. On Mon, Sep 29, 2014 at 4:11 PM, Deepthi Nandakumar < deep...@multicorewareinc.com> wrote: > Ashok/Santhoshini - pls review. Does removing offsets affect any planned >

Re: [x265] [PATCH] Changes for loadCTUData

2014-09-29 Thread Ashok Kumar Mishra
As of now we have not used offset value as per our plan. Later when we need, can add offset value. On Mon, Sep 29, 2014 at 4:46 PM, Santhoshini Sekar < santhosh...@multicorewareinc.com> wrote: > As of now offset can be removed from CU structure. We don't have any > planned optimization with it. I

Re: [x265] [PATCH] blockfill_s_16x16 avx2 asm code, performance improved 389.21 cycles -> 204.38 cycles

2014-09-29 Thread chen
At 2014-09-29 16:47:45,prav...@multicorewareinc.com wrote: ># HG changeset patch ># User Praveen Tiwari ># Date 1411980445 -19800 ># Node ID 9a8552ea378500baa21b89b24d8aec99acf7cce2 ># Parent 32f50df7fa7672f4c1818ddf3165b4bd243e0b10 >blockfill_s_16x16 avx2 asm code, performance improved 389.21

Re: [x265] [PATCH] blockfill_s_32x32 avx2 asm code, performance improved 1354.05 cycles -> 705.81 cycles

2014-09-29 Thread Steve Borho
On 09/29, prav...@multicorewareinc.com wrote: > # HG changeset patch > # User Praveen Tiwari > # Date 1411982102 -19800 > # Node ID 543a79749978dc4aae2956788bb16e50d2ceca14 > # Parent 9a8552ea378500baa21b89b24d8aec99acf7cce2 > blockfill_s_32x32 avx2 asm code, performance improved 1354.05 cycles ->

Re: [x265] [PATCH 0 of 2 ] TComDataCU: replace with more CU structure details

2014-09-29 Thread Steve Borho
This series looks ok, but unfortunately David's patches modify many of the same files and I wasn't able to rebase your changes on top of his without a considerable amount of merge conflicts. Can you resend these rebased on the new tip? -- Steve Borho __

[x265] sao: remove frame-based SAO

2014-09-29 Thread Satoshi Nakagawa
# HG changeset patch # User Satoshi Nakagawa # Date 1412038092 -32400 # Tue Sep 30 09:48:12 2014 +0900 # Node ID 3eacdaa304400b0100dcf1d1515ae1d24cbf4305 # Parent 5a6845566d1492d29af29ecc0cf75d644994735c sao: remove frame-based SAO diff -r 5a6845566d14 -r 3eacdaa30440 source/common/common.h

Re: [x265] sao: remove frame-based SAO

2014-09-29 Thread Steve Borho
On 09/30, Satoshi Nakagawa wrote: > # HG changeset patch > # User Satoshi Nakagawa > # Date 1412038092 -32400 > # Tue Sep 30 09:48:12 2014 +0900 > # Node ID 3eacdaa304400b0100dcf1d1515ae1d24cbf4305 > # Parent 5a6845566d1492d29af29ecc0cf75d644994735c > sao: remove frame-based SAO Queued toge

[x265] [PATCH 1 of 2] TComDataCU: replace getZorderIdxInCU() with encodeIdx of CU structure

2014-09-29 Thread santhoshini
# HG changeset patch # User Santhoshini Sekar # Date 1412047376 -19800 # Tue Sep 30 08:52:56 2014 +0530 # Node ID 21b1e8daa7e97e3828dfd948ff776951b939f423 # Parent 5a6845566d1492d29af29ecc0cf75d644994735c TComDataCU: replace getZorderIdxInCU() with encodeIdx of CU structure diff -r 5a684556

[x265] [PATCH 0 of 2 ] TComDataCU: replace with more CU structure details

2014-09-29 Thread santhoshini
___ x265-devel mailing list x265-devel@videolan.org https://mailman.videolan.org/listinfo/x265-devel

[x265] [PATCH 2 of 2] TComDataCU: replace getTotalNumPart() with CU structure details

2014-09-29 Thread santhoshini
# HG changeset patch # User Santhoshini Sekar # Date 1412047862 -19800 # Tue Sep 30 09:01:02 2014 +0530 # Node ID ecd15b4dfdeab9ddd16b11b1894aebfc66738c47 # Parent 21b1e8daa7e97e3828dfd948ff776951b939f423 TComDataCU: replace getTotalNumPart() with CU structure details diff -r 21b1e8daa7e9 -

[x265] [PATCH] blockfill_s_32x32 avx2 asm code: performance improved from 1354.05 cycles to 705.81 cycles, over sse version of asm code

2014-09-29 Thread praveen
# HG changeset patch # User Praveen Tiwari # Date 1412055614 -19800 # Node ID fb46cfd8ee611339da6af16f3c0de426eca3628f # Parent 975078c41433328fa35913ad46a95cd9c78a8bb2 blockfill_s_32x32 avx2 asm code: performance improved from 1354.05 cycles to 705.81 cycles, over sse version of asm code diff -

[x265] [PATCH] blockfill_s_16x16 avx2 asm code: performance improved from 389.21 cycles to 204.38 cycles, over sse version of asm code

2014-09-29 Thread praveen
# HG changeset patch # User Praveen Tiwari # Date 1412054832 -19800 # Node ID 975078c41433328fa35913ad46a95cd9c78a8bb2 # Parent 5a6845566d1492d29af29ecc0cf75d644994735c blockfill_s_16x16 avx2 asm code: performance improved from 389.21 cycles to 204.38 cycles, over sse version of asm code diff -r