Commit Graph

27 Commits

Author SHA1 Message Date
Glenn Randers-Pehrson
0a2d496d58 [libpng16] Imported from libpng-1.6.25rc03.tar 2016-08-29 16:10:33 -05:00
Glenn Randers-Pehrson
4be85e1b37 [libpng16] Rebased contrib/intel/intel_sse.patch 2016-08-17 06:52:07 -05:00
Glenn Randers-Pehrson
b50d5cea2b [libpng16] Imported from libpng-1.6.24.tar 2016-08-03 21:32:26 -05:00
Glenn Randers-Pehrson
96db0fff50 [libpng16] Conditionally compile ARM_NEON headers in pngpriv.h
Updated contrib/intel/intel_sse.patch
2016-08-02 09:31:05 -05:00
Glenn Randers-Pehrson
2c0b7a8ff6 [libpng16] Conditionally compile SSE2 headers in contrib/intel/intel_sse.patch 2016-07-31 21:22:06 -05:00
Glenn Randers-Pehrson
0a5e99d7ee [libpng16] Rebased contrib/intel/intel_sse.patch 2016-07-24 19:39:43 -05:00
Glenn Randers-Pehrson
6c7c5a04b8 [libpng16] More efficient absolute value calculation on SSE2 (Matthieu Darbois). 2016-06-20 08:28:34 -05:00
Matt Sarett
36762ac4b4 More efficient absolute value on SSE2 2016-06-20 09:06:25 -04:00
Glenn Randers-Pehrson
f757d74898 [libpng16] Moved sse2 prototype from pngpriv.h to contrib/intel/intel_sse.patch. 2016-05-31 09:45:17 -05:00
Glenn Randers-Pehrson
ac43a87820 [libpng16] Imported from libpng-1.6.22.tar 2016-05-26 07:25:14 -05:00
Glenn Randers-Pehrson
0794fee913 [libpng16] Imported from libpng-1.6.22beta06.tar 2016-05-05 19:37:54 -05:00
Glenn Randers-Pehrson
2b9f68631c [libpng16] Rebased contrib/intel_sse.patch. 2016-04-27 22:40:39 -05:00
Glenn Randers-Pehrson
eefd9127ca [libpng16] Imported from libpng-1.6.22beta05.tar 2016-04-27 13:01:14 -05:00
Glenn Randers-Pehrson
a967818235 [libpng16] Moved INTEL-SSE code from pngpriv.h to contrib/intel/intel_sse.patch. 2016-04-19 08:23:10 -05:00
Glenn Randers-Pehrson
5765a22249 [libpng16] Eliminate PNG_NO_INTEL_SSE_3BPP from intel INSTALL 2016-04-05 12:33:17 -05:00
Glenn Randers-Pehrson
f47e1d4995 [libpng16] Eliminated PNG_NO_INTEL_SSE_3BPP which was just for testing. 2016-04-05 12:16:27 -05:00
Mike Klein
8f506d3257 SSE filter speed improvements for bpp=3.
- memcpy-free implementations of load3() / store3().
    These should have less variance compiler to compiler.

    - call load3() only when needed at the end of a scanline.
    In the middle, we can use the faster load4(), ignoring that byte.
2016-04-04 16:10:09 -04:00
Glenn Randers-Pehrson
1a3fa1e091 [libpng16] Imported from libpng-1.6.22beta03.tar 2016-03-09 12:04:40 -06:00
Glenn Randers-Pehrson
e2d113cadf [libpng16] Updated INSTALL instructions for intel SSE code 2016-02-23 09:29:08 -06:00
Glenn Randers-Pehrson
b4f9f5d041 [libpng16] Fixed cut-and-paste errors in the new intel_sse.patch
and added PNG_NO_INTEL_SSE_3BPP macro.
2016-02-22 18:50:59 -06:00
Glenn Randers-Pehrson
6c3d5bd2a0 [libpng16] Moved contrib/intel/*.patch into INSTALL and added intel_sse.patch 2016-02-22 16:38:10 -06:00
Glenn Randers-Pehrson
da9d1d7aa6 [libpng16] Added pngdebug() statements to the new intel code and the arm code. 2016-02-19 14:58:59 -06:00
Glenn Randers-Pehrson
52846504da [libpng16] Fix copyright in sse code, use C-style comments 2016-02-19 09:46:51 -06:00
Glenn Randers-Pehrson
d06d66e899 [libpng16] Added contrib/intel/INSTALL 2016-02-18 22:17:46 -06:00
Glenn Randers-Pehrson
cdd9a23fd4 [libpng16] Fixed include directives in two new intel files to reflect
their new location in contrib/intel/* (was intel/*)
2016-02-18 21:23:24 -06:00
Matt Sarett
c3a45944e2 Fix paths in Makefile.am.patch 2016-02-18 12:45:13 -05:00
Matt Sarett
342c4eab2a Move sse opts into contrib/intel 2016-02-18 12:43:50 -05:00