hashirama/uvg266
1
0
Fork 0
mirror of https://github.com/ultravideo/uvg266.git synced 2024-11-24 10:34:05 +00:00
Code Issues Actions Projects Releases Wiki Activity
3317 commits 3 branches 6 tags 9.2 MiB
Commit graph

2812 commits

Author SHA1 Message Date
Jaakko Laitinen 8bd1a2b667 Update help message 2020-03-31 13:19:05 +03:00
Jaakko Laitinen b4f5486190 Set intra qp offset default to auto 2020-03-31 12:58:40 +03:00
Jaakko Laitinen 740688c67d Add auto option to intra qp offset 2020-03-31 11:56:44 +03:00
Pauli Oikkonen 0c7bfa7dc9 Fix AVX2 on Clang 
Besides just -mavx2, AVX2 support depends on a couple minor instruction
set extensions that should always exist on AVX2-capable hardware. Too
bad the different bit twiddling instructions are invoked slightly
differently between GCC and Clang, but now Clang seems to also produce
an AVX2-capable build.
 2020-03-26 18:48:48 +02:00
siivonek 89d3e674ce Comment out code which possible messes up OBA 2020-03-26 17:49:31 +02:00
siivonek be7d9ddec5 Fix error in frame variance calculation. Chroma channels were not added to variance 2020-03-26 14:33:00 +02:00
Jaakko Laitinen 45ca8f8113 Merge branch 'master' into 'extended_pu-depths' 2020-03-25 15:11:08 +02:00
siivonek 5986e71535 Fix mistake 2020-03-20 13:43:44 +02:00
Jaakko Laitinen d6ffe9e495 Update docs 2020-03-20 13:27:07 +02:00
Jaakko Laitinen 621450cc1d Update --help 2020-03-20 13:07:48 +02:00
Jaakko Laitinen aaac3df69b Add prefix to kvazaar.h define 2020-03-20 09:04:00 +02:00
siivonek 2a85be5752 Move qp_to_lambda so it is defined before use. Change some tabs to spaces 2020-03-19 22:13:53 +02:00
siivonek 0a4ce3c0aa Add vaq to new rate control 2020-03-19 21:43:52 +02:00
siivonek 1bbc598d75 Merge branch 'master' into vaq 2020-03-19 20:19:43 +02:00
Joose Sainio b53911d637 Merge branch 'rc-intra' 2020-03-19 13:34:15 +02:00
Joose Sainio a304a8ea6e Add weights for GOP 16 based on fitting a power curve to bits spent by HM 2020-03-19 11:13:43 +02:00
Joose Sainio e823ac1dae miscellaneous fixes 
- bump library version
- add help desk for --clip-neighbour
- update the default values of --clip-neighbour and --intra-bits
- update tests to more sensible
 2020-03-19 10:47:28 +02:00
Jaakko Laitinen b2ddba38c2 Set correct size for pu-depth min/max data structure 2020-03-19 09:29:43 +02:00
Joose Sainio 2c345bc3cf try to fix tsan issue 2020-03-18 14:58:54 +02:00
Jaakko Laitinen fe428dcbe1 Fix no gop functionality 2020-03-18 11:03:33 +02:00
Jaakko Laitinen af3d559d8d Let pu-depth be defined per gop-layer 2020-03-17 17:57:18 +02:00
Ari Lemmetti cbd77944d8 Costs in rough intra search may be negative. Get rid of UBSan error. 2020-03-16 22:13:14 +02:00
Ari Lemmetti aa0ade3f65 Cast values to unsigned to make UBSan not trigger due to left-shifting negatives 2020-03-16 19:52:34 +02:00
RLamm 27fe716654 Fixed reference POC indexing 2020-03-11 15:33:37 +02:00
RLamm bf24831780 Attempt to fix random crashes 2020-03-11 15:31:47 +02:00
RLamm 887659db1f Attempted to scale the extra_mvs 2020-03-11 15:31:46 +02:00
siivonek 8d9719ff90 Merge branch 'master' into vaq 2020-03-05 14:17:01 +02:00
Joose Sainio c9a8f2a596 Completely disable intra based model for frame 1 2020-03-04 12:52:13 +02:00
Joose Sainio 19c79c3e58 don't use the intra frame based estimation if the result is bad 2020-03-04 09:26:22 +02:00
Ari Lemmetti 7b7358c25a Update presets veryslow and placebo a bit 
Both use now --gop 16, --intra-qp-offset -3, --me tz, and --transform-skip
 2020-03-03 20:41:01 +02:00
Pauli Oikkonen 60e7956dc5 Disable inaccurate integer variance calculation for now 2020-03-02 19:18:55 +02:00
Pauli Oikkonen fc1b91335b Implement variance calculation in integer math 
Maybe this is a bit faster than FP, it's not accurate though
 2020-03-02 18:17:18 +02:00
Pauli Oikkonen 35c825c75f Move hsum_8x32b to avx2_common_functions 2020-02-27 17:52:17 +02:00
Pauli Oikkonen b00ac7d1c4 AVX2 version of buffer variance calculation 2020-02-25 15:57:56 +02:00
siivonek a380e43bda Add chroma channels to variance calculation. 2020-02-24 19:54:34 +02:00
Pauli Oikkonen 1bd9c6dd93 Make a strategy out of pixel_var 2020-02-24 19:37:36 +02:00
Pauli Oikkonen 86ebf366e1 fix typo 2020-02-24 18:18:10 +02:00
Joose Sainio f81de41775 Merge branch 'master' into rc-intra 2020-02-24 15:30:57 +02:00
siivonek 5688bcd646 Merge branch 'master' into vaq 2020-02-21 17:11:10 +02:00
siivonek 908ecb1767 Add rounding to aq offsets. Fix typo 2020-02-21 13:51:43 +02:00
Ari Lemmetti 1dfc69b42e Consider merge index bits in merge analysis and early skip 2020-02-20 09:43:58 +02:00
Joose Sainio 7deb22c8e8 Merge branch 'master' into rc-intra 2020-02-19 15:01:04 +02:00
Kari Siivonen (TAU) c972ca9067 Add assert to check if deltaQP out of bounds. Clip adaptive QP to [-13, 12]. 2020-02-18 13:20:26 +02:00
Kari Siivonen (TAU) f07990794f Fix error in vaq pixel blit range calculation 2020-02-18 13:20:26 +02:00
Kari Siivonen (TAU) 57ed40c263 Fix application of aq offset 2020-02-18 13:20:26 +02:00
Kari Siivonen (TAU) be2f420d61 Change: vaq requires parameter. Parameter defines vaq strength ex. 15 == 1.5 2020-02-18 13:20:26 +02:00
Kari Siivonen (TAU) bf1b2c1e22 Add define for vaq strength parameter 2020-02-18 13:20:26 +02:00
Kari Siivonen (TAU) 150559a7e8 Fix bugs. Enable set_qp_in_cu when using vaq 2020-02-18 13:20:26 +02:00
Kari Siivonen (TAU) c8c71274ee Change tabs to spaces. 2020-02-18 13:20:26 +02:00
siivonek 888382953d Implement calculation of vaq values. Values not used yet. 2020-02-18 13:20:25 +02:00
siivonek ad40a88c09 Add no-vaq option to vaq 2020-02-18 13:20:25 +02:00
siivonek 09f0a1c52e Fix typo in comment 2020-02-18 13:20:25 +02:00
siivonek 84fb3fd7d1 aq: Add --vaq commandline option 2020-02-18 13:20:25 +02:00
Joose Sainio 2a98f5db1e fix intra-bits for lp-gop 2020-02-18 10:38:29 +02:00
Ari Lemmetti 71d9327f62 Further improve fast bipred 2020-02-17 20:32:52 +02:00
Ari Lemmetti 80c26870d5 Update docs 2020-02-15 23:29:18 +02:00
Ari Lemmetti ebb183cc01 Add option to make intra QP offset configurable 2020-02-15 22:54:48 +02:00
Ari Lemmetti be3e08d6db Add gop.h to Makefile 2020-02-15 22:54:47 +02:00
Ari Lemmetti 1354acd358 Prevent negative values being written to SPS with --gop=0 2020-02-15 22:54:47 +02:00
Ari Lemmetti fe4869916c Disable GOP and intra qp offset for all-intra coding automatically 2020-02-15 22:54:46 +02:00
Ari Lemmetti 9849fb7c77 Enable experimental rate control for GOP 16 2020-02-15 22:54:46 +02:00
Ari Lemmetti a0a22dec8a Remove deprecated / unused lambda adjustments 2020-02-15 22:54:46 +02:00
Arttu Ylä-Outinen 829a70e6a7 Copy lowdelay GOP definition from HM 2020-02-15 22:36:58 +02:00
Arttu Ylä-Outinen 28f99c0b87 Change definition of 8-GOP to match HM 2020-02-15 22:36:58 +02:00
Arttu Ylä-Outinen 636fa8fbdd Fix maximum decoded picture buffer size 2020-02-15 22:36:57 +02:00
Arttu Ylä-Outinen ebd5156db5 Add definition for random access GOP of length 16 2020-02-15 22:36:57 +02:00
Arttu Ylä-Outinen 6653f06dd0 Only compute GOP layer weights when RC is enabled 2020-02-15 22:36:57 +02:00
Arttu Ylä-Outinen c8fff1e0d6 Use a larger number of bits for POC lsb when needed 
Changes the number of bits used for coding the least significant bits of
the POC based on the GOP size.
 2020-02-15 22:36:56 +02:00
Arttu Ylä-Outinen d757a832c2 Change GOP QP offset handling to match HM 
Adds fields qp_model_scale and qp_model_offset to kvz_gop_config and
intra_qp_offset to kvz_config.
 2020-02-15 22:36:56 +02:00
Arttu Ylä-Outinen f37dcd5879 Move GOP definition to a separate file 
Moves definition of the 8-GOP from cfg.c to gop.h.
 2020-02-15 22:36:55 +02:00
Ari Lemmetti 6e1007a3e7 Get rid of LAMBA! (Commit #3000) 2020-02-15 22:32:52 +02:00
Ari Lemmetti 0c02e71b43 Remove minor error from readme 2020-02-15 22:29:08 +02:00
Joose Sainio e90d3141a2 Merge branch 'master' into rc-intra 2020-02-05 11:06:56 +02:00
Ari Lemmetti 9a0236bb4e Add option 'zero-coeff-rdo' 2020-02-04 21:26:29 +02:00
Ari Lemmetti 886ff36d12 Initial implementation of fast bipred. 2020-02-04 15:46:23 +02:00
Ari Lemmetti 3c7dd0752f Remove the broken "no mov" branch. 
Causes hash mismatches for example in SlideShow sequence.
 2020-02-03 15:26:31 +02:00
RLamm bf8941ddb8 Added comment about partial-coding usage 2020-01-31 16:19:48 +02:00
RLamm b8488ab48d Changed "partial-coding" variables to uint32_t 2020-01-31 16:02:29 +02:00
RLamm 76e3249754 Changed parameter "slicer" to "partial-coding" to avoid confusion. 2020-01-31 14:22:32 +02:00
RLamm 30d5df40c5 Custom headers for the distributed coding 2020-01-29 15:54:49 +02:00
Joose Sainio 54571529a4 Fix accessing previous frame that didn't exist 2020-01-17 10:48:35 +02:00
Joose Sainio 5c671d20e1 Use the new clipping only in situations where it actually helps 2020-01-17 09:08:21 +02:00
Joose Sainio 3c34d7c863 Fix qp estimation and checking of previous frames that dont exist 2020-01-15 09:32:04 +02:00
Joose Sainio 1a35c22a52 Change clipping of lambda and qp for ctus on OBA rc 
instead of clipping qp and lambda to the value of last value from the state
clip to previous frame with same layer and if such frame doesn't exist, clip
to previous frame
 2020-01-14 14:46:05 +02:00
Pauli Oikkonen c3d9e97e9f Fix VS build 2019-12-12 18:34:55 +02:00
Pauli Oikkonen 7f238ca299 Remove debug print functions 
Whoops
 2019-12-12 18:19:31 +02:00
Pauli Oikkonen eefb5e50b3 De-inline pred_filtered_dc functions, shouldn't make much difference though 2019-12-12 17:30:00 +02:00
Pauli Oikkonen 169314de4f 32x32 filtered DC prediction in AVX2 2019-12-11 18:17:06 +02:00
Pauli Oikkonen fb2481b7e4 16x16 filtered DC implemented in AVX2 2019-12-10 15:54:50 +02:00
Joose Sainio b78aa7b272 save c and k to frame 2019-12-06 10:52:54 +02:00
Joose Sainio 5b10e5fb7e parameterize the clipping option 2019-12-06 09:51:04 +02:00
Pauli Oikkonen da370ea36d Implement AVX2 8x8 filtered DC algorithm 2019-11-28 14:10:10 +02:00
Pauli Oikkonen 5d9b7019ca Implement a 4x4 filtered DC pred function 2019-11-26 17:05:54 +02:00
Joose Sainio ca0060cbba try the original clipping 2019-11-26 15:13:04 +02:00
Pauli Oikkonen f1485ab087 Start doing an arbitrary size filtered DC pred - maybe easier to just create separate functions for fixed block sizes? 2019-11-25 15:20:29 +02:00
Joose Sainio ab2fded8af Update threadwrapper to enable pthread_rwlock_t 2019-11-21 13:38:40 +02:00
Joose Sainio eb78aead1f Fix additional potential data races 2019-11-21 11:03:12 +02:00
Joose Sainio 35d7e0d88b Fix data race 2019-11-21 10:25:04 +02:00
Pauli Oikkonen 979d66031c Create a strategy out of intra_pred_filtered_dc 2019-11-19 14:50:31 +02:00
Joose Sainio 0e8815a3d8 test clipping qp to previous frame instead of previous ctus 2019-11-19 14:32:31 +02:00
Joose Sainio ddb4e5a131 move the intra bit calculation so that it is used also with lambda rc 2019-11-19 14:16:48 +02:00
Joose Sainio a07833f3e6 check that mallocs in rc initialization were successful 
only call kvz_update_after_picture when using the OBA rc
 2019-11-19 13:59:44 +02:00
Joose Sainio 50d410a316 re-enable static qp encoding and lambda rc 2019-11-19 13:45:58 +02:00
Pauli Oikkonen fa4bb86406 Optimize intra_pred_planar_avx2 for 4x4 blocks 2019-11-19 13:39:02 +02:00
Joose Sainio 57e5615ece Fix incorrect intra rc calculation skipping 2019-11-19 13:25:31 +02:00
Joose Sainio 6cc3bcd87e Command line parameters for oba rc and implementation of the usage of the intra parameter 2019-11-19 09:29:06 +02:00
Joose Sainio eb73548af5 Encode first frame completely before starting others to enable owf 2019-11-18 09:51:37 +02:00
Pauli Oikkonen 4761d228f9 Start to vectorize the 4x4 loop 2019-11-15 17:32:40 +02:00
Pauli Oikkonen 8d45ab4951 Stupidify the 4x4 planar loop for vectorization 2019-11-14 17:14:04 +02:00
Joose Sainio c759c138ed Prepare the rc data structure to be shared among all frame encoders 2019-11-13 11:56:25 +02:00
Joose Sainio cdb7c851a4 Fix weight calculation 2019-11-13 08:55:31 +02:00
Joose Sainio b9b01f8036 WPP with threading 2019-11-12 12:12:57 +02:00
Joose Sainio 615973adca should enable threading with wpp when owf is not used 2019-11-12 09:03:00 +02:00
Pauli Oikkonen 6f13f6525c Merge branch 'new_prints' 2019-11-07 17:04:21 +02:00
Joose Sainio d353f7dd1a Disable debug prints, fix multiple bugs in the calculation 2019-11-07 15:08:57 +02:00
mercat 57e8c3ebc2 Merge branch 'ML-cplx_red_ICIP' 2019-11-07 13:25:47 +02:00
Pauli Oikkonen 558f0ec401 Mbps, not mbps 2019-11-05 18:06:00 +02:00
Pauli Oikkonen 2edf533925 Tidy the end report printing 
Also fix a bug with non-integer target FPS
 2019-11-05 17:20:00 +02:00
Joose Sainio 408fd4ccb6 Fix lambda and qp calcualtion for intra frames 
also fixes a bug with selecting the clip neighbor lambda and clip neighbor qp
selection for inter frames
 2019-11-05 10:51:39 +02:00
Pauli Oikkonen c7313ce567 Store AVG QP information in encmain 2019-11-04 17:08:07 +02:00
Reima Hyvönen 80575c59bf Some updates done to get right bitrate and avg QP 2019-10-31 15:56:24 +02:00
Reima Hyvönen 252bab8820 Added prints to bitrate and AVG QP 2019-10-31 15:56:24 +02:00
Pauli Oikkonen 6d7a4f555c Also remove 16x16 (A * B^T)^T matrix multiply 
Can be done using (B * A^T) instead, it's the exact same
 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 2c2deb2366 Tidy AVX2 32x32 matrix multiply 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 98ad78b333 Tidy the old AVX2 32x32 matrix multiply 
It was actually a very good algorithm, just looked messy!
 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 4a921cbdb5 Retain data as much in YMM registers as possible 
This seems to make it a whole lot quicker
 2019-10-28 16:19:42 +02:00
Pauli Oikkonen ac4d710e23 Unroll 32x32 matrix multiply, use all regs 2019-10-28 16:19:42 +02:00
Pauli Oikkonen a58608d0b8 Remove totally unnecessary (A * B^T)^T 32x32 multiply 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 043f53539f Implement a streamlined matrix-multiply 32x32 DCT 2019-10-28 16:19:42 +02:00
Pauli Oikkonen e9da2d851b Tidy 32x32 fast DCT's helper functions 2019-10-28 16:19:42 +02:00
Pauli Oikkonen e382339182 Implement fast (butterfly) 32x32 DCT in AVX2 2019-10-28 16:19:42 +02:00
Pauli Oikkonen b5962dadac Tidy indentation in AVX2 16x16 iDCT 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 36a8f89025 Fine-tune 16x16 AVX2 iDCT 2019-10-28 16:19:42 +02:00
Pauli Oikkonen ca9409de2b Implement 16x16 DCT as butterfly algorithm in AVX2 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 7c69a26717 Use aligned loads and stores for AVX2 DCT 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 8e9c65dca6 Align DCT matrices and temp transform buffers 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 148a150522 Align DCT source and dest blocks to cache line 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 8e60bbf6a6 Slightly tune 16x16 forward DCT 
Use an array of __m256i's to store temporary value, essentially letting
the compiler enforce alignment and use aligned loads and stores.
 2019-10-28 16:19:42 +02:00
Pauli Oikkonen c0cc0e8a75 Optimize 16x16 multiply by only slicing right mat once 2019-10-28 16:19:42 +02:00
Pauli Oikkonen e463d27f22 Implement streamlined generic 16x16 matrix multiply 
It can't be this fast for real, can it?
 2019-10-28 16:19:42 +02:00
Pauli Oikkonen beb85ce9d6 Reorder parameters for 8x8 matrix multiplies 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 292af62256 Implement tailored 16x16 forward DCT 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 30ce461d98 Redo 4x4 matrix multiplication 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 07970ea82f Streamline by-the-book 8x8 matrix multiplication 
Also chop up the forward transform into two tailored multiply functions
 2019-10-28 16:19:42 +02:00
Pauli Oikkonen 7ec7ab3361 Implement a tailored AVX2 8x8 DCT 2019-10-28 16:19:42 +02:00
Joose Sainio 372934c7db Fix division by zero 2019-10-10 16:35:56 +03:00
Joose Sainio 9bdfdeaf5c Rest of the owl 2019-10-09 15:48:58 +03:00
Joose Sainio 1ba8525faf WIP 2019-10-09 10:35:07 +03:00
Joose Sainio 19496d2692 ? 2019-10-03 14:50:11 +03:00
Joose Sainio 4b111e339e fix couple of bugs in the implementation, bit calculation seems still bit off 2019-10-01 15:08:39 +03:00
Joose Sainio 84615e406a fix compiler warnings 2019-09-27 14:20:08 +03:00
Joose Sainio 14b7a75713 Call the new functions and fix bugs 2019-09-27 14:14:24 +03:00
Joose Sainio ef74bfb182 unify naming 2019-09-27 10:16:21 +03:00
Joose Sainio e36f481bda qp calculation for frame 2019-09-27 09:05:40 +03:00
Joose Sainio 47019ca1cd intra ck update 2019-09-26 16:04:53 +03:00
Joose Sainio 7c8f4da7cb Update c and k except after first intra 2019-09-26 13:09:28 +03:00
Joose Sainio 0577d481c1 CTU level code 2019-09-25 12:12:21 +03:00
pkubaj 1d7fcf4227
Fix build on powerpc64 with LLVM 2019-09-12 15:05:00 +02:00
mercat 0de567bfa4 Fixe memory leak 2019-09-12 09:45:32 +03:00
mercat fa116de619 Add static 2019-09-11 16:18:12 +03:00
mercat b8753a9293 Fucking INLINE fixed 2019-09-11 16:12:07 +03:00
mercat b855144e68 INLINE fixe 2019-09-11 16:12:07 +03:00
mercat 694337b803 Add const and more const 2019-09-11 16:12:07 +03:00
mercat 21c07638ed Remove const into kvz_init_constraint. 2019-09-11 16:12:06 +03:00
mercat 2bca507abe Clean version of machine learning constraint code. (ICIP paper) 2019-09-11 16:12:06 +03:00
Alexandre Mercat 0f4b7be6ee First version of ML ICIP code for master 2019-09-11 16:12:06 +03:00
Pauli Oikkonen 99597b828a Work around the ancient Win32 calling convention hassle 
See if this'll work now
 2019-09-06 13:14:42 +03:00
Pauli Oikkonen c5ca18950c Revert "Revert to 6924d90052 due to broken visual studio build" 
This reverts commit 1dd0619bd7.
 2019-09-05 18:21:55 +03:00
Pauli Oikkonen 55529decd5 Implement _mm256_insert_epi32 and extract pseudo-ops 
Visual Studio headers apparently lack these guys
 2019-09-05 18:20:52 +03:00
Ari Lemmetti 147378e1f9 Prevent 8x4 and 4x8 bipred in merge analysis 2019-09-03 16:32:50 +03:00
Ari Lemmetti ef1fdbf259 Separate prediction of single PU/PB from CU/CB 2019-09-03 16:32:50 +03:00
Joose Sainio 7d2737bdf6 WIP picture lambda calculation 2019-09-03 11:03:35 +03:00
Ari Lemmetti 3bc510712f Enable merge analysis for smp and amp 2019-09-02 17:31:51 +03:00
Ari Lemmetti 557bcbc6aa Make luma or chroma only inter "recon" or predict possible 2019-09-02 17:15:28 +03:00
Joose Sainio 131c04f65c Fix incorrect weight for intra frame 2019-08-29 12:01:13 +03:00
Joose Sainio 8f96678d13 Fix issue with intra frames being part of gop when they shouldn't 2019-08-29 09:28:10 +03:00
Ari Lemmetti aa8ab195d1 Compare rough cost of the best merge mode against AMVP to make mode decision 2019-08-26 22:49:09 +03:00
Ari Lemmetti 8f866ff83a Use correct index 2019-08-26 20:10:10 +03:00
Ari Lemmetti 2343958a14 Fix transform split for small luma blocks 2019-08-24 21:50:17 +03:00
Ari Lemmetti 800fc8644d Reset CBFs because CBFs might have been set earlier for depth earlier. 2019-08-24 21:49:33 +03:00
Ari Lemmetti a80de22bc7 Add only different candidates to the list 2019-08-24 21:49:33 +03:00
Ari Lemmetti 45c7961412 Remove tr depth fill. It should not be needed. 2019-08-24 21:49:32 +03:00
Ari Lemmetti ff8711aaab Add missing logic to add valid indices to list 2019-08-24 21:49:29 +03:00
Ari Lemmetti 1dd0619bd7 Revert to 6924d90052 due to broken visual studio build 2019-08-08 15:15:34 +03:00
Pauli Oikkonen 2852baa673 Separate sign3_diff_epu8 from calc_eo_cat 
Just to keep things simple, clear and obvious
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 17947b79ee Add sao_shared_generics.h in Makefile.am 2019-08-07 16:35:24 +03:00
Pauli Oikkonen a8dd6ce351 Add a note about having implemented a separate AVX2 version of SAO offset array calculation 2019-08-07 16:35:24 +03:00
Pauli Oikkonen a858e7dd4b Combine duplicate code into inline functions 2019-08-07 16:35:24 +03:00
Pauli Oikkonen de0e97f711 Take 8/16/24b loads and stores into separate functions 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 10979f58fe Tidy up code 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 9cc11976c0 Combine the delta accumulation from edge and band ddistortion into shared func 
This won't reduce object size, but there'll be less duplicate code
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 55d877bd66 Vectorize sao_edge_ddistortion 2019-08-07 16:35:24 +03:00
Pauli Oikkonen aef0f301d3 Fix function signatures 
Mark anything intended as read-only to be const, and fix alignment
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 997fd369b3 Redo calc_sao_edge_dir_avx2 
Do it wider, 32 pixels at once!
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen db1e475e02 Use i32 instead of i8 for x/y offsets 
Doesn't matter too much, because this number isn't used in SIMD
computation, only as a memory reference offset.
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 12de466ef5 Reimplement non-band SAO color reconstruction in AVX2 
Streamline things to work on 32 pixels at once instead of 8
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen e8bff99329 Redo the SAO_TYPE_BAND subsection of AVX2 SAO color reconstruction 
Vectorize it all, hope this helps with perf
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 7b5dffa855 Implement calc_sao_offset_array in AVX2 
To be efficient, the AVX2 color reconstruction algorithm will need
offsets in byte, not dword, arrays. This is completely specific to 8-bit
pixels and the function signature is fundamentally distinct from the
generic algorithm, so it's better to not strategize SAO offset array
calculation.
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 29563b7039 Make kvz_calc_sao_offset_array more obvious 
Name temporary values from array lookups etc that are referred multiple
times to, to make the behavior of the mechanism more transparent. Define
all the constant values at the beginning of the function and declare as
const.
 2019-08-07 16:35:24 +03:00
Pauli Oikkonen 08881f5e9b (TEMP) (TODO) (whatever) Avoid compiler warnings 
I want the CI to not crash on its -Wall -Werror, but instead to actually
build the thing and report me about actual memory errors etc
 2019-08-07 16:35:24 +03:00