Commit graph

  • 2811ce58f4 [avx2] AVX2 version of depquant now exactly matches scalar version Joose Sainio 2023-04-22 11:53:54 +0300
  • 48ea4bff4d [dep_quant] Fix rate_estimator and quant_block init cases Joose Sainio 2023-04-21 11:34:31 +0300
  • dfff9a8030 [avx2] Move dep quant stuff to strategies Joose Sainio 2023-04-17 15:14:35 +0300
  • 0591342b3a [avx2] replace or Joose Sainio 2023-04-19 14:00:36 +0300
  • 8b1d6fab59 [avx2] Replace loads and stores with non-avx512 stores Joose Sainio 2023-04-19 13:07:38 +0300
  • 6d0a3fa5fc [avx2] Replace _mm_and_epi32 with _mm_and_si128 Joose Sainio 2023-04-19 12:34:43 +0300
  • 7fdc045690 [dep_quant] Clean up Joose Sainio 2023-04-19 12:01:24 +0300
  • 8eb0f66734 [depquant] update_state_eos_avx2 working Joose Sainio 2023-04-18 15:43:30 +0300
  • 00cc58bc55 [depquant] Only initialize rate_estimator when necessary Joose Sainio 2023-04-17 14:45:55 +0300
  • 00f838306f [depquant] Initialize quant_block only when necessary Joose Sainio 2023-04-17 14:18:57 +0300
  • 9e27b4056a [avx2] WIP update_state_eos_avx2 Joose Sainio 2023-04-17 13:52:42 +0300
  • c56350b8d6 [avx2] and last Joose Sainio 2023-04-14 09:55:09 +0300
  • 9f69713c24 [depquant] remove an unnecessary memcpy Joose Sainio 2023-04-14 08:25:33 +0300
  • aa48943c22 [avx2] Do decision cost comparison with avx2 Joose Sainio 2023-04-12 15:36:45 +0300
  • cd6110cfac [depquant] Pre calculate things Joose Sainio 2023-04-12 15:02:06 +0300
  • 8f4c3cecbf [avx2] update_states_avx2 working Joose Sainio 2023-04-12 10:41:37 +0300
  • 58a66c0654 [avx2] WIP update_states_avx2 Joose Sainio 2023-04-10 15:31:05 +0300
  • 04be92a8ec [avx2] simplify Joose Sainio 2023-04-10 08:40:35 +0300
  • 8b19c468cf [avx2] check_rd_costs_avx2 done Joose Sainio 2023-04-09 14:05:50 +0300
  • c6e6f5da33 [avx2] WIP check_rd_costs_avx2, almost? Joose Sainio 2023-04-08 18:58:40 +0300
  • 8caabcde1a [avx2] WIP check_rd_costs_avx2 Joose Sainio 2023-04-07 15:15:58 +0300
  • 2912db5fca [dep_quant.c] Small refactor Joose Sainio 2023-04-07 08:14:06 +0300
  • 64d34f8559 [depquant] AoS -> SoA pre quant Joose Sainio 2023-04-06 16:07:38 +0300
  • 2f1e9c4020 [depquant] AoS -> SoA all states Joose Sainio 2023-04-06 15:12:29 +0300
  • 73442f1bba [depquant] AoS -> SoA for Decision Joose Sainio 2023-04-05 11:17:16 +0300
  • 26ef1dda09 [ibc] Fix chroma SAD handling and disable chroma SAD for now Marko Viitanen 2023-08-30 15:06:08 +0300
  • bcdd339a04 Add some comments. avx2-dct siivonek 2023-08-22 14:01:58 +0300
  • 8d743be0d4 [avx2] Fix issue with 16x32 inverse transform Joose Sainio 2023-07-27 12:23:32 +0300
  • 6cff22589e static all the things Joose Sainio 2023-07-26 15:24:31 +0300
  • c295d4c855 Fix compilation errors Joose Sainio 2023-07-26 15:20:33 +0300
  • 30e8e3874a Mostly working Joose Sainio 2023-07-26 14:05:04 +0300
  • 07d6b45489 [avx2] Inverses work when ISP is not enabled Joose Sainio 2023-07-26 10:45:39 +0300
  • 9f8837fa9c [avx2] Forward transforms seem to be working siivonek 2023-07-24 15:32:53 +0300
  • 7a8256176e Disable all avx2 optimizations that cannot be used with mtt/isp Joose Sainio 2023-07-21 14:23:37 +0300
  • 0239572796 [ibc] Fix some instances where CU_INTER was checked instead of !CU_INTRA Marko Viitanen 2023-08-23 15:21:45 +0300
  • 312ac6731c [ibc] dual-tree rebase fixes Marko Viitanen 2023-08-15 13:24:22 +0300
  • 805afb1331 [fix] Minor fixes Joose Sainio 2023-04-05 09:33:00 +0300
  • 8c14fa94ba [mtt] Fix small issues with luma and chroma searches Joose Sainio 2023-03-02 15:31:34 +0200
  • 7a5245c5a4 [dual-tree] Fix chroma tree split model context derivation during search Joose Sainio 2023-02-23 13:41:59 +0200
  • 707e11dbcf [dual-tree] Small fixes Joose Sainio 2023-02-23 12:50:30 +0200
  • 91591c7e7c [dual-tree] Remove the limitation of not allowing 2 height chroma blocks in dual tree Joose Sainio 2023-02-23 08:48:08 +0200
  • 146e1cb85e [dual-tree] WIP simplification Joose Sainio 2023-02-22 14:48:00 +0200
  • 0f50caa2d0 [mtt] Fix various small issues and DepQuant for non-square blocks Joose Sainio 2023-02-15 14:23:55 +0200
  • d222718c22 [mtt] Minor fixes Joose Sainio 2023-02-14 10:36:01 +0200
  • d69bdf79f4 [mtt] Fix couple of issues with 64x32 CUs and non square tr skip rdoq Joose Sainio 2023-02-08 14:39:36 +0200
  • 7d787c6b22 [ISP] Fix ISP cost calculation and DepQuant with mts Joose Sainio 2023-02-03 14:58:36 +0200
  • 6e24b9a7f9 [DepQuant] Fix isp+depquant and trskip + isp Joose Sainio 2023-01-31 14:26:00 +0200
  • 93c7e9c296 [DepQuant] Fix for mts and lfnst being quantized incorrectly during search Joose Sainio 2023-01-23 13:39:22 +0200
  • dc652c75f9 [DepQuant] Isp and chroma Joose Sainio 2023-01-19 16:30:47 +0200
  • 505c26eef3 [DepQuant] Fix Joose Sainio 2023-01-17 11:03:14 +0200
  • c6087230a8 [DepQuant] Fix Joose Sainio 2023-01-17 11:03:14 +0200
  • 5abe9e57c6 [DepQuant] Working but not necessarily improving Joose Sainio 2023-01-12 15:05:37 +0200
  • 5236bc93be [DepQuant] WIP: doesn't crash but bitstream is illegal and quality a lot worse Joose Sainio 2023-01-11 14:17:18 +0200
  • bfa699fac6 [DepQuant] WPP: API Joose Sainio 2023-01-11 10:12:59 +0200
  • f8994a7fae [DepQuant] WIP: dequant Joose Sainio 2023-01-11 09:25:34 +0200
  • 3e66a897d4 [DepQuant] WIP: compiles Joose Sainio 2023-01-10 15:32:07 +0200
  • 4dbe0cd6c3 [DepQuant] WIP: easy part done Joose Sainio 2023-01-09 14:10:10 +0200
  • 2a33af283e [DepQuant] WIP: initialization done Joose Sainio 2023-01-05 14:21:43 +0200
  • 1373a7ac1d [mtt] correct indexing for chroma tree Joose Sainio 2023-01-04 11:19:01 +0200
  • d3f42949a7 [mtt] Only consider termination if the cu is completely inside the frame Joose Sainio 2023-01-04 09:55:22 +0200
  • 0c63743fc0 [mtt] Early terminations for all intra Joose Sainio 2023-01-04 09:34:34 +0200
  • bd3ec75173 [mtt] search early terminations Joose Sainio 2023-01-02 13:47:40 +0200
  • 2d00cab4b9 [isp] properly reset cabac context during intra search Joose Sainio 2022-12-21 14:33:10 +0200
  • b27eca7c37 [deblock] fix width and height to correct order Joose Sainio 2022-12-21 13:45:56 +0200
  • eae7d72384 [isp] Keep cabac contexts up to date for the different isp tus Joose Sainio 2022-12-21 11:53:55 +0200
  • c744f79117 [mtt] Fix rdoq for non-square blocks Joose Sainio 2022-12-21 09:17:55 +0200
  • 3b09c66d25 [deblock] Use the isp block dimensions instead of cu dimensions fro deblock Joose Sainio 2022-12-20 11:25:58 +0200
  • 73956a9a46 [isp] Fix isp bitcost calculation Joose Sainio 2022-12-20 09:13:08 +0200
  • f3c8a4f5db [lfnst] Also chroma can only use lfnst if dimensions are minimum 4 Joose Sainio 2022-12-20 08:23:01 +0200
  • a36a1fb5ff [mtt] There is always at least the height or width amount reference pixels available Joose Sainio 2022-12-19 14:42:45 +0200
  • af23c81afa [mtt] Fix reading uninitialized data for local chroma tree Joose Sainio 2022-12-19 14:25:03 +0200
  • 9acdab3209 [mtt] Fix lfnst bit counting for 64 wide or tall chroma tree cus Joose Sainio 2022-12-19 13:22:10 +0200
  • 812377db45 [mtt] Set cus outside of the frame to zero for initializing partial worktree Joose Sainio 2022-12-19 10:24:00 +0200
  • 27d114bc08 [mtt] Fix negative indexing Joose Sainio 2022-12-19 09:59:33 +0200
  • 4e203108bc [mtt] Fix ref pixel generation for the second half of 32x2 chroma cus Joose Sainio 2022-12-19 09:35:32 +0200
  • 446c53fd00 [mtt] Fix cclm for non 64 divisible heights Joose Sainio 2022-12-16 14:13:45 +0200
  • ad2bb20f23 [mtt] Fix deblock for isp and properly set the limit for cclm Joose Sainio 2022-12-16 09:37:56 +0200
  • c89ebf8bf1 [cclm] Fix heap corruption for non 64 divisible frames Joose Sainio 2022-12-15 13:12:42 +0200
  • d296cac7c3 [mtt] fix reference building for 16x1 Joose Sainio 2022-12-15 12:41:12 +0200
  • 926ed7e145 [rdoq] partly fix rdoq for 16x1 and 1x16 Joose Sainio 2022-12-15 11:25:12 +0200
  • 8e4b864e6b [deblock] Fix incorrect direction for transform split of tall blocks at the top CTU row also for chroma Joose Sainio 2022-12-15 10:07:47 +0200
  • 34aed10ec1 [mtt] fix Joose Sainio 2022-12-15 09:09:38 +0200
  • 1333ab55d9 [mtt] Fix ref building for 32x64 cus Joose Sainio 2022-12-14 13:12:03 +0200
  • 1493a2616c [mtt] fix getting collocated chroma for edge cus Joose Sainio 2022-12-14 12:33:24 +0200
  • ffe17e48d7 [mtt] minor fixes Joose Sainio 2022-12-14 12:02:00 +0200
  • 74591cd39b [tests] no-cpuid for test_cabac_state.sh since it now has mtt in it Joose Sainio 2022-12-14 11:38:30 +0200
  • 972670a177 [tests] chmod +x Joose Sainio 2022-12-14 11:35:44 +0200
  • 5aa13ad62a [tests] add tests for mtt Joose Sainio 2022-12-14 11:26:01 +0200
  • 06fa86c340 [isp] Fix coordinates Joose Sainio 2022-12-14 10:54:52 +0200
  • 71516b8155 [mtt] Make sure mtt splits cannot reach a situation where search cannot be performed Joose Sainio 2022-12-14 10:41:36 +0200
  • 23e6b9f56c [mtt] Check that we are inside the CTU before checking the ctu data Joose Sainio 2022-12-14 10:23:20 +0200
  • facbc794bf [mtt] Fix trying to get split data from depth -1 Joose Sainio 2022-12-14 10:16:51 +0200
  • 567fa7b2bd [deblock] Fix incorrect direction for transform split of tall blocks at the top CTU row Joose Sainio 2022-12-14 10:10:08 +0200
  • 9c2574880a [mtt] Fix deblock for --combine-intra Joose Sainio 2022-12-14 09:25:54 +0200
  • 90ce1390c0 [mtt] static Joose Sainio 2022-12-14 09:11:40 +0200
  • f6ecb15ced [mtt] Fix implicit splits when mtt is not enabled Joose Sainio 2022-12-14 09:01:52 +0200
  • 05218bae21 [jccr] jccr=4 hasn't been necessary for a long time Joose Sainio 2022-12-14 08:47:24 +0200
  • b69e9b2958 [mtt] Fix final issues? Joose Sainio 2022-12-14 08:37:06 +0200
  • 6620ba8d76 [mtt] fix deblock Joose Sainio 2022-12-13 14:51:38 +0200
  • 09baddef17 [mtt] Fix lfnst and chroma coeffs and tests Joose Sainio 2022-12-12 10:05:17 +0200