Johann d6a7489dd5 neon variance: process two rows of 8 at a time ...

When the width is equal to 8, process two rows at a time. This doubles
the speed of 8x4 and improves 8x8 by about 20%.

8x16 was using this technique already, but still improved a little bit
with the rewrite.

Also use this for vpx_get8x8var_neon

BUG=webm:1422

Change-Id: Id602909afcec683665536d11298b7387ac0a1207

2017-05-04 08:59:46 -07:00

..

arm

neon variance: process two rows of 8 at a time

2017-05-04 08:59:46 -07:00

mips

Fix mips msa fwd xform mismatch

2017-03-22 14:01:03 +05:30

ppc

ppc: Add convolve8_vsx and convolve8_avg_vsx

2017-05-02 20:27:47 -07:00

x86

Update highbd convolve functions arguments to use uint16_t src/dst

2017-04-25 14:22:19 -07:00

add_noise.c

add_noise,vpx_setup_noise: correct 'char_dist' type

2016-08-24 21:46:26 -07:00

avg.c

Fix real-time compression regression in hbd mode

2017-01-31 23:17:09 -08:00

bitreader_buffer.c

remove CONFIG_MISC_FIXES

2017-03-24 20:41:39 -07:00

bitreader_buffer.h

vp10: code sign bit before absolute value in non-arithcoded header.

2015-09-16 19:35:03 -04:00

bitreader.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

bitreader.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

bitwriter_buffer.c

remove CONFIG_MISC_FIXES

2017-03-24 20:41:39 -07:00

bitwriter_buffer.h

vp10: code sign bit before absolute value in non-arithcoded header.

2015-09-16 19:35:03 -04:00

bitwriter.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

bitwriter.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

deblock.c

deblock: annotate postproc parameters

2017-01-31 15:58:57 -08:00

fastssim.c

Add compiler warning flag -Wextra and fix related warnings.

2016-09-27 12:05:01 -07:00

fwd_txfm.c

Fix warnings reported by -Wshadow: Part1: vpx_dsp directory

2016-10-17 19:25:19 -07:00

fwd_txfm.h

Replace vp9_ prefix in 2D-DCT functions with vpx_

2015-07-28 16:06:44 -07:00

intrapred.c

intrapred: sync highbd_d135_predictor w/d135_

2017-03-24 20:45:44 -07:00

inv_txfm.c

Add vpx_highbd_idct32x32_135_add_c()

2017-03-08 10:46:33 -08:00

inv_txfm.h

Replace highbd_dct_const_round_shift with dct_const_round_shift

2016-11-04 16:15:02 -07:00

loopfilter.c

[vpx highbd lpf NEON 1/6] horizontal 4

2016-10-06 14:03:04 -07:00

postproc.h

postproc : fix function parameters for noise functions.

2016-07-15 08:27:34 -07:00

prob.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

prob.h

get_prob(): rationalize int types

2017-02-24 15:36:52 -08:00

psnr.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

psnr.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

psnrhvs.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

quantize.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

quantize.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

sad.c

sad avg: align intermediate buffer

2017-04-17 14:26:33 +00:00

ssim.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

ssim.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

subtract.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

sum_squares.c

Support measure distortion in the pixel domain

2016-07-06 18:25:17 -07:00

txfm_common.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

variance.c

re-enable vpx_comp_avg_pred_sse2

2017-04-17 08:40:43 -07:00

variance.h

Remove halfpix specialization

2016-08-23 17:05:39 -07:00

vpx_convolve.c

Update highbd convolve functions arguments to use uint16_t src/dst

2017-04-25 14:22:19 -07:00

vpx_convolve.h

Update highbd convolve functions arguments to use uint16_t src/dst

2017-04-25 14:22:19 -07:00

vpx_dsp_common.h

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

vpx_dsp_rtcd_defs.pl

neon variance: add small missing sizes

2017-05-04 08:59:42 -07:00

vpx_dsp_rtcd.c

vpx_dsp: apply clang-format

2016-07-25 14:14:19 -07:00

vpx_dsp.mk

ppc: Add convolve_copy

2017-04-29 15:47:25 +02:00

vpx_filter.h

add vpx high bitdepth convolve8 NEON intrinsics optimization

2016-10-17 15:23:54 -07:00