]>
granicus.if.org Git - libvpx/log
Dmitry Kovalev [Tue, 26 Nov 2013 23:35:11 +0000 (15:35 -0800)]
Removing dummy assignments.
Change-Id: I10d1a4bcac751a982d9dd135f019e3a4d92f8522
Yaowu Xu [Tue, 26 Nov 2013 22:32:02 +0000 (14:32 -0800)]
Merge "Amended some comments for clarity"
Dmitry Kovalev [Tue, 26 Nov 2013 22:13:30 +0000 (14:13 -0800)]
Merge "Deleting vp9_treereader.h file."
Dmitry Kovalev [Tue, 26 Nov 2013 20:38:58 +0000 (12:38 -0800)]
Deleting vp9_treereader.h file.
Renaming treed_read() to consistent vp9_read_tree() and moving it from
deleted vp9_treereader.h to vp9_dboolhuff.h file.
Change-Id: Iedd8655acbe25e4fcf62b79e5a13bdea69b6b004
Yunqing Wang [Tue, 26 Nov 2013 19:56:13 +0000 (11:56 -0800)]
Merge "Add a noise test vector"
James Zern [Tue, 26 Nov 2013 19:49:01 +0000 (11:49 -0800)]
Merge "thumb: Fix a typo in a comment"
Yunqing Wang [Tue, 26 Nov 2013 01:02:31 +0000 (17:02 -0800)]
Add a noise test vector
Added the test vector provided by Attila, which caught the bug in
Issue 661 "Decoder produces mismatched outputs with ssse3 enabled
and disabled"
vp90-hantro-stream-001.ivf
size: 320x180; 20 frames
Change-Id: Ic0d2b57ac7596ecb938dd55abc8c706fc2dd6d8f
Yaowu Xu [Tue, 26 Nov 2013 19:22:30 +0000 (11:22 -0800)]
Merge "Fix unit test failures"
Frank Galligan [Tue, 26 Nov 2013 00:47:14 +0000 (16:47 -0800)]
Fix 16 wide neon horz loopfilter.
Multiply by 3 was on 8bit vectors when it should have been on
16bit vectors.
Change-Id: I248c1429b3134dfd171dfab0ebb109fd2437e1fc
Yaowu Xu [Tue, 26 Nov 2013 17:56:14 +0000 (09:56 -0800)]
Fix unit test failures
Change-Id: Ibc61ef81fafeb20df6df6e5496b6c01760f3dc84
Yunqing Wang [Tue, 26 Nov 2013 17:35:14 +0000 (09:35 -0800)]
Merge "Do vertical loopfiltering in parallel"
Deb Mukherjee [Tue, 26 Nov 2013 17:34:20 +0000 (09:34 -0800)]
Merge "Some cleanups on rate control"
Martin Storsjo [Tue, 26 Nov 2013 08:56:43 +0000 (10:56 +0200)]
thumb: Fix a typo in a comment
Change-Id: Ic318be9d74a95793abc41501fe563164de53a6d7
Dmitry Kovalev [Tue, 26 Nov 2013 03:15:21 +0000 (19:15 -0800)]
Merge "Moving mv entropy encodings calculation to the encoder side."
Dmitry Kovalev [Tue, 26 Nov 2013 03:14:47 +0000 (19:14 -0800)]
Merge "Removing vp9_modecosts.{c, h} files."
Deb Mukherjee [Tue, 26 Nov 2013 02:58:45 +0000 (18:58 -0800)]
Some cleanups on rate control
Removes the active_worst_qchanged variable since it is never
set to 1.
Change-Id: I29a291fd1068fd9b504a2db7768d45644c1eae3e
Dmitry Kovalev [Tue, 26 Nov 2013 00:36:28 +0000 (16:36 -0800)]
Moving mv entropy encodings calculation to the encoder side.
Moved arrays:
vp9_mv_joint_encodings
vp9_mv_class_encodings
vp9_mv_class0_encodings
vp9_mv_fp_encodings
Change-Id: Iaf5008c579fcbd6d77fdd81d1aef8c71b5f308b7
Dmitry Kovalev [Tue, 26 Nov 2013 00:08:42 +0000 (16:08 -0800)]
Merge "Removing redundant call of vp9_init_mbmode_probs()."
Dmitry Kovalev [Mon, 25 Nov 2013 20:44:05 +0000 (12:44 -0800)]
Removing vp9_modecosts.{c, h} files.
Renaming vp9_init_mode_costs() to fill_mode_costs() and moving it to
vp9_rdopt.c.
Change-Id: Ib2542d216458f6dced9f4b7ccbdd2cd98176aa5a
Tom Finegan [Mon, 25 Nov 2013 20:05:19 +0000 (12:05 -0800)]
vpxenc: Move config check code into its own source files.
- Add command line args that allow display of warnings without prompting
for user input.
- Extend warning code to make it somewhat scalable.
Change-Id: I2bad8f9315f6eed120c2e1bbe0a2a5ede15fbf35
Dmitry Kovalev [Fri, 22 Nov 2013 19:45:30 +0000 (11:45 -0800)]
Reusing txfrm_block_to_raster_xy in vp9_xform_quant().
Change-Id: Ib273dfff3be284f3c9ae288e5315fb6c6126f9c2
Dmitry Kovalev [Mon, 25 Nov 2013 19:03:00 +0000 (11:03 -0800)]
Merge "Cleaning up vp9_write_nmv_probs() function."
Dmitry Kovalev [Mon, 25 Nov 2013 18:59:56 +0000 (10:59 -0800)]
Merge "Adding select_tx_size() function."
Dmitry Kovalev [Mon, 25 Nov 2013 18:59:24 +0000 (10:59 -0800)]
Merge "Moving {left, right}_block_mode to vp9_blockd.h."
Dmitry Kovalev [Mon, 25 Nov 2013 18:59:08 +0000 (10:59 -0800)]
Merge "Renaming COMPPREDMODE_TYPE enum and its members."
Yaowu Xu [Mon, 25 Nov 2013 17:58:25 +0000 (09:58 -0800)]
Amended some comments for clarity
Change-Id: I31c3908ba394095deb5d3a5d7b7c9b2b5328c3e8
Yaowu Xu [Mon, 25 Nov 2013 18:52:22 +0000 (10:52 -0800)]
Merge "In frame Q adjustment experiment."
Jingning Han [Mon, 25 Nov 2013 18:29:07 +0000 (10:29 -0800)]
Merge "Use separate inter predictors for enc/dec"
Paul Wilkins [Fri, 22 Nov 2013 17:21:53 +0000 (17:21 +0000)]
In frame Q adjustment experiment.
The idea here is to allow "in frame" adjustment of the final Q
value used to encode each SB64, using segmentation.
There is also adjustment of the rd mult in regions of overspend.
Activated using aq_mode=2
Change-Id: I2f140cd898c9f877c32cd6d2e667f5e11ada4b1c
Yaowu Xu [Mon, 25 Nov 2013 18:20:53 +0000 (10:20 -0800)]
Merge "Fix a build issue with visual c."
Guillaume Martres [Mon, 25 Nov 2013 17:45:55 +0000 (09:45 -0800)]
Merge "Rename block_yrd_txfm to block_rd_txfm"
Jingning Han [Fri, 22 Nov 2013 19:56:29 +0000 (11:56 -0800)]
Use separate inter predictors for enc/dec
The decoder will construct inter predictor using lazy border extension,
while the encoder, going with multiple runs of motion search in the rate-
distortion optimization loop for each block, does border extension at
frame level. This commit makes separate the inter predictors for encoder
and decoder, respectively.
Change-Id: Ieca2fecba3a7201a6d64ef9f219e5d91e50559c3
Jingning Han [Mon, 25 Nov 2013 17:14:46 +0000 (09:14 -0800)]
Merge "Separate setup_scale_factor/extend_frame_borders"
Yaowu Xu [Mon, 25 Nov 2013 16:11:35 +0000 (08:11 -0800)]
Fix a build issue with visual c.
Change-Id: Ic8fc16ee1734cfde0d12a2e3abb3e9299382f3b1
Dmitry Kovalev [Mon, 25 Nov 2013 00:20:59 +0000 (16:20 -0800)]
Cleaning up vp9_write_nmv_probs() function.
Change-Id: I9af38f1a0e226ee094c17081f044a722ac1caded
Guillaume Martres [Mon, 25 Nov 2013 00:14:54 +0000 (01:14 +0100)]
Rename block_yrd_txfm to block_rd_txfm
This function is also used with the U and V planes.
Change-Id: Ib6beae0438790bfb690b3a2bda9c5c8b0e89a9b8
Dmitry Kovalev [Mon, 25 Nov 2013 00:08:29 +0000 (16:08 -0800)]
Merge "Using partition counts from FRAME_COUNTS struct in the encoder."
Dmitry Kovalev [Mon, 25 Nov 2013 00:08:20 +0000 (16:08 -0800)]
Merge "Cleaning up vp9_append_sub8x8_mvs_for_idx."
Dmitry Kovalev [Mon, 25 Nov 2013 00:07:57 +0000 (16:07 -0800)]
Merge "Inlining and removing vp9_set_pred_flag_seg_id() function."
Paul Wilkins [Sat, 23 Nov 2013 11:05:55 +0000 (03:05 -0800)]
Merge "Added cpuid compatibility for older MSVC versions"
Dmitry Kovalev [Sat, 23 Nov 2013 01:32:11 +0000 (17:32 -0800)]
Inlining and removing vp9_set_pred_flag_seg_id() function.
Change-Id: I0fd76937e847f78378a7ab3fa0af00a7c2c52b42
Dmitry Kovalev [Sat, 23 Nov 2013 01:16:19 +0000 (17:16 -0800)]
Merge "Organizing all scan tables into lookup table."
Yaowu Xu [Sat, 23 Nov 2013 00:45:56 +0000 (16:45 -0800)]
Added cpuid compatibility for older MSVC versions
Change-Id: I891bf936e03411ca611620e7cb2eb5081993a346
Yaowu Xu [Sat, 23 Nov 2013 01:01:07 +0000 (17:01 -0800)]
Merge "Fix bug in extend_frame chroma extended too far"
Dmitry Kovalev [Sat, 23 Nov 2013 00:35:37 +0000 (16:35 -0800)]
Renaming COMPPREDMODE_TYPE enum and its members.
List of renames:
COMPPREDMODE_TYPE => REFERENCE_MODE
SINGLE_PREDICTION_ONLY => SINGLE_REFERENCE
COMP_PREDICTION_ONLY => COMPOUND_REFERENCE
HYBRID_PREDICTION => REFERENCE_MODE_SELECT (like TX_MODE_SELECT)
NB_PREDICTION_TYPES => REFERENCE_MODES
Change-Id: If723dabe9435325d0165dcd028142a2c78b417b4
Dmitry Kovalev [Sat, 23 Nov 2013 00:20:45 +0000 (16:20 -0800)]
Organizing all scan tables into lookup table.
Change-Id: Ie829ee58a55157e6972c63cebe69a5d0a3221349
Dmitry Kovalev [Fri, 22 Nov 2013 23:28:32 +0000 (15:28 -0800)]
Cleaning up vp9_append_sub8x8_mvs_for_idx.
Change-Id: Ic92f15d82ff5cfa3df655d08e460335c2ef8a325
Dmitry Kovalev [Fri, 22 Nov 2013 22:26:39 +0000 (14:26 -0800)]
Using partition counts from FRAME_COUNTS struct in the encoder.
Change-Id: I6c3d47b00acabe7ffba22ffc73741173aa9a0bff
Jingning Han [Fri, 22 Nov 2013 18:56:41 +0000 (10:56 -0800)]
Separate setup_scale_factor/extend_frame_borders
This commit takes out vp9_extend_frame_borders from
vp9_setup_scale_factors.
The refactoring is for the preparation of the use of lazy border
extension at decoder. This makes it necessary to handle border
extension separately at encoder/decoder. The use of
vp9_extend_frame_borders will be removed, when lazy border extension
is ready.
Change-Id: Ia3baba3d179d5f11eee1634f19b3b319d2a59186
Adrian Grange [Tue, 19 Nov 2013 22:01:44 +0000 (14:01 -0800)]
Fix decoder to handle display size correctly
The decoder ignored the display width & height
specified in the frame header.
This patch adds a control, VP9D_GET_DISPLAY_SIZE, to
allow the application to obtain the display width and
height from the frame header.
vpxdec has been modified to scale the output frame to
this size.
Should the request for the display size fail vpxdec will
use the native width and height of the raw decoded
frame instead.
Change-Id: I25db04407426dac730263720c75a7dd6400af68a
Dmitry Kovalev [Fri, 22 Nov 2013 18:52:40 +0000 (10:52 -0800)]
Merge "Cleaning up entropy probability update in encoder."
Dmitry Kovalev [Fri, 22 Nov 2013 18:51:38 +0000 (10:51 -0800)]
Merge "Removing txfrm_block_to_raster_xy() call from extend_for_intra()."
Yunqing Wang [Fri, 22 Nov 2013 18:39:55 +0000 (10:39 -0800)]
Merge "Improve vp9_fdct4x4_sse2 (x1.2)"
Yunqing Wang [Fri, 22 Nov 2013 00:43:37 +0000 (16:43 -0800)]
Do vertical loopfiltering in parallel
This patch followed "Add filter_selectively_vert_row2 to enable
parallel loopfiltering" commit, and added x86 SSE2 optimization
to do 16-pixel filtering in parallel. For other optimizations
(neon and dspr2), current 16-pixel functions were done by calling
8-pixel functions twice, and real 16-pixel functions could be added
later.
Decoder speedup:
tulip clip: 2% speed gain;
old_town_cross: 1.2% speed gain;
bus: 2% speed gain.
Change-Id: I4818a0c72f84b34f5fe678e496cf4a10238574b7
Yaowu Xu [Fri, 22 Nov 2013 18:03:51 +0000 (10:03 -0800)]
Merge "Fix the cpuid macro for x86_64 non-gcc build"
Adrian Grange [Fri, 22 Nov 2013 01:19:04 +0000 (17:19 -0800)]
Fix bug in extend_frame chroma extended too far
This fixes issue 667.
In the case where the frame was an odd number of pixels
wide or high, the border was being extended by one col
or row too far.
The calculation of color plane dimensions was modified
to use those already computed at the time the frame
buffer was allocated.
Also freed the temporary scaling buffer in vpxdec to
prevent a memory leak.
Change-Id: Ied04bdcdfd77469731408c05da205db1a6f89bf5
Jim Bankoski [Fri, 22 Nov 2013 16:16:17 +0000 (08:16 -0800)]
Merge changes Id1698a35,Idcabd0b9
* changes:
detokenization speedups
Don't write 0's to token_cache
Deb Mukherjee [Fri, 22 Nov 2013 16:06:48 +0000 (08:06 -0800)]
Merge "Refactoring of rate control - part 1"
Deb Mukherjee [Wed, 6 Nov 2013 21:13:59 +0000 (13:13 -0800)]
Refactoring of rate control - part 1
Moves all rate control variables to a separate structure,
removes some currently unused variables,
moves some rate control functions to vp9_ratectrl.c,
and splits the encode_frame_to_data_rate function.
Change-Id: I4ed54c24764b3b6de2dd676484f01473724ab52b
Dmitry Kovalev [Fri, 22 Nov 2013 03:30:58 +0000 (19:30 -0800)]
Removing txfrm_block_to_raster_xy() call from extend_for_intra().
Change-Id: I6a48d1f35ed5fe7a2c7499675b339994c9c3bdf2
Yaowu Xu [Fri, 22 Nov 2013 01:39:33 +0000 (17:39 -0800)]
Fix the cpuid macro for x86_64 non-gcc build
Change-Id: I0c44800db10db8d74c1ddfe89abecfd1c53d0f8d
Tom Finegan [Fri, 22 Nov 2013 01:56:26 +0000 (17:56 -0800)]
Merge "vpxenc: Add vpxenc.h and move/rename the global_config struct"
Jim Bankoski [Fri, 22 Nov 2013 00:55:22 +0000 (16:55 -0800)]
detokenization speedups
removed unnecessary ifs and branches ..
Change-Id: Id1698a35292659388f48926791024d1400f2cea9
Dmitry Kovalev [Fri, 22 Nov 2013 00:48:34 +0000 (16:48 -0800)]
Merge "Using num_4x4_blocks_* instead of b_{width, height}_log2."
Tom Finegan [Fri, 22 Nov 2013 00:46:40 +0000 (16:46 -0800)]
vpxenc: Add vpxenc.h and move/rename the global_config struct
- Rename the struct to VpxEncoderConfig.
- The idea behind this is to enable checking the global settings against
stream specific settings in source files other than vpxenc.c.
Change-Id: Ic736cbb714845b9466acb34671780d65b83ad1a8
Dmitry Kovalev [Fri, 22 Nov 2013 00:37:27 +0000 (16:37 -0800)]
Merge "Removing plane_block_{width, height} functions."
Dmitry Kovalev [Fri, 22 Nov 2013 00:24:22 +0000 (16:24 -0800)]
Merge "Using txfrm_block_to_raster_xy() in encoder."
Dmitry Kovalev [Thu, 21 Nov 2013 23:55:40 +0000 (15:55 -0800)]
Adding select_tx_size() function.
Change-Id: I9d18b31661a2ccdcd4e25956882c7fc2d4b7002e
Dmitry Kovalev [Thu, 21 Nov 2013 23:53:06 +0000 (15:53 -0800)]
Using num_4x4_blocks_* instead of b_{width, height}_log2.
Change-Id: I9ea3946c17b19f511565cd771037abe7db8b3ddb
Joshua Litt [Thu, 21 Nov 2013 23:06:51 +0000 (15:06 -0800)]
Merge "Removing PARAMS macro for consistency"
Frank Galligan [Thu, 21 Nov 2013 23:06:17 +0000 (15:06 -0800)]
Merge "Revert "Add 16 wide neon horz loopfilter.""
Frank Galligan [Thu, 21 Nov 2013 22:01:33 +0000 (14:01 -0800)]
Revert "Add 16 wide neon horz loopfilter."
The change caused mismatches with some test vectors on neon.
Original CL: https://gerrit.chromium.org/gerrit/#/c/67863/
Change-Id: I913891636d53783e93cb1865ca78ded1821dc4b0
Jim Bankoski [Thu, 21 Nov 2013 20:52:15 +0000 (12:52 -0800)]
Don't write 0's to token_cache
This code only updates the token_cache if the result is non0.
Change-Id: Idcabd0b993a926fea9c29dbec134b9c5c4859b40
Dmitry Kovalev [Thu, 21 Nov 2013 20:36:02 +0000 (12:36 -0800)]
Syncing update_coef_probs() implementation with decoder.
Using for loop based on max_tx_size instead of separate checks. Combining
build_coeff_contexts() with update_coef_probs().
Change-Id: Ie335a7db29830677fbc14478a9c190d3c1068665
Abo Talib Mahfoodh [Thu, 21 Nov 2013 20:00:20 +0000 (15:00 -0500)]
Improve vp9_fdct4x4_sse2 (x1.2)
Modifications are done to reduce the total clock cycle.
Speedup: 1.2
Tested with: park_joy_420_720p50.y4m
Change-Id: Ia36b87e62e2f80a5fadaf5628729aedc80f38f3f
Dmitry Kovalev [Thu, 21 Nov 2013 19:43:53 +0000 (11:43 -0800)]
Moving {left, right}_block_mode to vp9_blockd.h.
Both functions have no relation to motion vectors, so moving them from
vp9_findnearmv.h to vp9_blockd.h.
Change-Id: I74f524267886ab0fff4a2da793a10c906ed0f43a
Yunqing Wang [Thu, 21 Nov 2013 19:25:55 +0000 (11:25 -0800)]
Merge "Add filter_selectively_vert_row2 to enable parallel loopfiltering"
hkuang [Thu, 21 Nov 2013 19:24:02 +0000 (11:24 -0800)]
Merge "Remove unnecessary eob checking."
Frank Galligan [Thu, 21 Nov 2013 18:29:30 +0000 (10:29 -0800)]
Merge "Add 16 wide neon horz loopfilter."
Yunqing Wang [Fri, 15 Nov 2013 19:04:09 +0000 (11:04 -0800)]
Add filter_selectively_vert_row2 to enable parallel loopfiltering
Added filter_selectively_vert_row2 to be ready for parallel
loopfiltering in vertical direction. This change did 2-row
filtering at a time. If 2 vertically adjacent 8x8 blocks do same
type of filtering, we can do 16-pixel filtering in parallel.
Next, we need to provide 16-pixel loopfiltering functions in c
and optimized versions for codec speedup.
Change-Id: Idf97bbdd70566e55bd30e1fd25cb8544e33291be
Yunqing Wang [Thu, 21 Nov 2013 17:40:02 +0000 (09:40 -0800)]
Merge "Correct ssse3 8/16-pixel wide sub-pixel filter calculation"
Frank Galligan [Tue, 19 Nov 2013 18:33:32 +0000 (10:33 -0800)]
Add 16 wide neon horz loopfilter.
Add support to do 16 pixel horizontal filtering in Neon.
Nexus devices saw about 0.5% decode speed increase.
Change-Id: I2993f6c2d49f31fa74976879eeaa289fd3f4e15d
Dmitry Kovalev [Thu, 21 Nov 2013 01:14:40 +0000 (17:14 -0800)]
Removing redundant call of vp9_init_mbmode_probs().
This function is called from vp9_setup_past_independence() which is called
before the modified piece of code. Moving reset of inter_mode_probs into
vp9_init_mbmode_probs() for consistency.
Change-Id: Ib188e8798e1fbe15407fd501406761b746fdda95
Tom Finegan [Thu, 21 Nov 2013 01:18:28 +0000 (17:18 -0800)]
vpxenc: Warn users about incorrect quantizer settings.
Also, clean up stylistically questionable code near my changes.
Change-Id: I92c96a274cb339b7b74174a608f94ae86aba8354
Dmitry Kovalev [Wed, 20 Nov 2013 22:43:03 +0000 (14:43 -0800)]
Merge "Removing old code."
Dmitry Kovalev [Wed, 20 Nov 2013 22:39:58 +0000 (14:39 -0800)]
Merge "Adding MV_FP_SIZE constant."
Dmitry Kovalev [Wed, 20 Nov 2013 22:39:50 +0000 (14:39 -0800)]
Merge "Using is_inter_block() and has_second_ref() functions."
Dmitry Kovalev [Wed, 20 Nov 2013 22:05:21 +0000 (14:05 -0800)]
Removing old code.
Change-Id: I67d1681c7b17661deb792c5e6a9e2014a73ff9b7
Dmitry Kovalev [Wed, 20 Nov 2013 21:58:21 +0000 (13:58 -0800)]
Using txfrm_block_to_raster_xy() in encoder.
Change-Id: Ibe847000467fe46bf8ce87d8f1ef8f2d5ad1eaf4
Yunqing Wang [Wed, 20 Nov 2013 20:52:56 +0000 (12:52 -0800)]
Correct ssse3 8/16-pixel wide sub-pixel filter calculation
Although no mismatch was indicated for 8/16 wide sub-pixel filters
in issue 661, they had similar problems that could cause mismatch
potentially. This patch fixed calculations in HORIZx8/16
and VERTx8/16.
Change-Id: I169961c9d40a20340995b7d22aafc89ccf30bfca
Dmitry Kovalev [Wed, 20 Nov 2013 20:39:29 +0000 (12:39 -0800)]
Removing plane_block_{width, height} functions.
Change-Id: I29c0dfcf41a1253d5e2a0d2ff740c0c38ebaa5a2
Jim Bankoski [Wed, 20 Nov 2013 20:30:03 +0000 (12:30 -0800)]
Merge "Clean up removal of vp9_pareto8 table."
Dmitry Kovalev [Wed, 20 Nov 2013 04:25:55 +0000 (20:25 -0800)]
Using is_inter_block() and has_second_ref() functions.
Change-Id: Iadd771a33c8874f3b774923bca4da3c8fe5429ee
Dmitry Kovalev [Wed, 20 Nov 2013 04:18:01 +0000 (20:18 -0800)]
Adding MV_FP_SIZE constant.
Change-Id: I98d750ee92ff51fb714980418ea28be3b1d0f3c6
Yunqing Wang [Wed, 20 Nov 2013 20:01:39 +0000 (12:01 -0800)]
Merge "Support for extended feature flags enumeration leaf in CPUID instruction"
hkuang [Wed, 20 Nov 2013 19:22:00 +0000 (11:22 -0800)]
Remove unnecessary eob checking.
Change-Id: Ia568f70bddc1a2b62141a0197459119ca74c22b5
Jim Bankoski [Wed, 20 Nov 2013 19:34:30 +0000 (11:34 -0800)]
Merge "remove the model and copy in pack_mb_tokens"
Jim Bankoski [Wed, 20 Nov 2013 19:17:26 +0000 (11:17 -0800)]
Clean up removal of vp9_pareto8 table.
Change-Id: I5556e8d1fc150be8a3e93af21900829b59a500dc
Erik Niemeyer [Wed, 20 Nov 2013 04:11:57 +0000 (21:11 -0700)]
Support for extended feature flags enumeration leaf in CPUID instruction
This CL fixes an overcite with the AVX2 support CL previously
merged (Change-Id: Idc03f3fca4bf2d0afd33631ea1d3caf8fc34ec29) that
prevented runtime execution of AVX2 code in WebM.
Background:
Starting with the Sandybridge processor, the CPUID instruction was
enhanced to add various extended feature flag enumeration leaves.
Reading these leaves requires an additional input value for the CPUID
instruction which is stored in ECX. This change adds this second input
value for all ARCH_X86 and ARCH_x86_64 targets to the CPUID macros,
allowing checks of EBX bit 5 for AVX2 support. This capability will be
required moving forward to check for future processor features.
Change-Id: Ie9d872bc9ff68dad4b6578e4544e4dfd0ae26c36
Jingning Han [Wed, 20 Nov 2013 18:55:27 +0000 (10:55 -0800)]
Merge "Take out assertion from inverse transforms"