13 Commits

Author SHA1 Message Date
Jingning Han
9fe31390ca Support rectangular tx_size in the common lib
Change-Id: I4128ab932a967a3d657bb1f95f0fa2af20a06469
2016-11-02 11:48:31 -07:00
Yi Luo
7317200002 Hybrid inverse transforms 16x16 AVX2 optimization
- Add unit tests to verify the bit-exact result.
- User level time reduction (EXT_TX):
    encoder: 3.63%
    decoder: 2.36%
- Also add tx_type=V_DCT...H_FLIPADST SSE2 for 16x16 inv txfm.

Change-Id: Idc6d9e8254aa536e5f18a87fa0d37c6bd551c083
2016-11-01 13:38:20 -07:00
Yi Luo
1a0f27aaa6 Fix avx2 16x16/32x32 fwd txfm coeff output on HBD
Change-Id: Ida036defe5688894a63007a31aa2dd0b3f0b5d59
2016-10-21 14:14:00 -07:00
Yi Luo
157e45a44b Fix the overflow of av1_fht32x32() in 2D DCT_DCT
- Use range check function to avoid DCT_DCT overflow.
  We need to re-develop the column txfm side scaling/rounding. Now,
  we prefer to maintain the current BDRate level.
- Encoder user level time reduction <1% owing to av1_fht32x32_avx2.
- Add MemCheck unit test and fdct32() unit test.

Change-Id: I1e67030f67bc637859798ebe2f6698afffb8531c
2016-10-20 09:22:24 -07:00
Yi Luo
e9fde265f7 Zero high 128b YMM registers to avoid SSE-AVX transition penalties
Documents:
- https://software.intel.com/en-us/articles/intel-avx-state-transitions-migrating-sse-code-to-avx
- https://software.intel.com/sites/default/files/m/d/4/1/d/8/11MC12_Avoiding_2BAVX-SSE_2BTransition_2BPenalties_2Brh_2Bfinal.pdf

Change-Id: I90f85fcb15a7a2c49ee068300be6ffe9c68d371c
2016-10-14 12:22:35 -07:00
David Barker
33231d4801 Add sse2 forward and inverse 16x32 and 32x16 transforms
Change-Id: I1241257430f1e08ead1ce0f31db8272b50783102
2016-10-13 14:01:22 +01:00
Yi Luo
fed8e1c06d Hybrid forward transform 32x32 AVX2 optimization
- av1_fht32x32 AVX2 function level time reduction ~89% compared to C.

- av1_fht32x32_avx2() on DCT_DCT improves 42.62% over aom_fdct32x32_avx2()
  But function replacement must go with the corresponding inverse txfm.

- No obvious user level time reduction due to 32x32 TX_TYPE selection.

- Zero high 128b YMM to avoid AVX-SSE transition penalties
  (fix 16x16 case).

- Added 32x32 AVX2 unit tests to verify bitexact.

- AVX2 optimization summary:
  On CPU i7-6700, based on 16x16/32x32 fwd txfm optimization results:
  C to AVX2: function level time reduction, ~86-89%.
  SSE2 to AVX2: function level time reduction, ~51%.

Change-Id: Idd0cd8bf066a61c7117140ef15ab6c1f8eb4b036
2016-10-12 14:19:53 -07:00
David Barker
4d03d6fc6f Add sse2 forward / inverse 4x8 and 8x4 transforms
Change-Id: I89ed93fb20cf975c2b463cff58879521ceaa4163
2016-10-10 09:02:45 -07:00
Yi Luo
e8e8cd8f1b Hybrid forward transforms 16x16 AVX2 optimization
- Unit tests are added for AVX2 SIMD.
- Encoder speed improvement:
  AV1 baseline and EXT_TX, three 1080p sequences at bitrate:
  800 Kbps, 2 Mbps, 6 Mbps, on i7-6700 CPU, average
  user level time reduction: 3.86%.

Change-Id: Ibbd7837ee3a831c6b1e4e471bf6c8d3fa3a19ff4
2016-10-06 15:33:15 -07:00
Geza Lore
1a800f6539 Add SSE2 versions of av1_fht8x16 and av1_fht16x8
Encoder speedup ~2% with ext-tx + rect-tx

Change-Id: Id56ddf102a887de31d181bde6d8ef8c4f03da945
2016-09-09 11:29:41 -07:00
Yaowu Xu
9c01aa1b0c Change to use aom copyright notice
This minimize code differences between AOM master and nextgenv2

Change-Id: If144865bdf3ef0818e7aac11018b9e786444c550
2016-09-02 08:22:07 -07:00
Yaowu Xu
f883b42cab Port renaming changes from AOMedia
Cherry-Picked the following commits:
0defd8f Changed "WebM" to "AOMedia" & "webm" to "aomedia"
54e6676 Replace "VPx" by "AVx"
5082a36 Change "Vpx" to "Avx"
7df44f1 Replace "Vp9" w/ "Av1"
967f722 Remove kVp9CodecId
828f30c Change "Vp8" to "AOM"
030b5ff AUTHORS regenerated
2524cae Add ref-mv experimental flag
016762b Change copyright notice to AOMedia form
81e5526 Replace vp9 w/ av1
9b94565 Add missing files
fa8ca9f Change "vp9" to "av1"
ec838b7  Convert "vp8" to "aom"
80edfa0 Change "VP9" to "AV1"
d1a11fb Change "vp8" to "aom"
7b58251 Point to WebM test data
dd1a5c8 Replace "VP8" with "AOM"
ff00fc0 Change "VPX" to "AOM"
01dee0b Change "vp10" to "av1" in source code
cebe6f0 Convert "vpx" to "aom"
17b0567 rename vp10*.mk to av1_*.mk
fe5f8a8 rename files vp10_* to av1_*

Change-Id: I6fc3d18eb11fc171e46140c836ad5339cf6c9419
2016-08-31 18:19:03 -07:00
Yaowu Xu
c27fc14b02 Port folder renaming changes from AOM
Manually cherry-picked commits:
ceef058 libvpx->libaom part2
3d26d91 libvpx -> libaom
cfea7dd vp10/ -> av1/
3a8eff7 Fix a build issue for a test
bf4202e Rename vpx to aom

Change-Id: I1b0eb5a40796e3aaf41c58984b4229a439a597dc
2016-08-31 17:26:24 -07:00