Commit Graph

9 Commits

Author SHA1 Message Date
Steve Snyder
6a4d539353 Bug 512865. qcms: Improve SSE2 performance, add SSE support. r=jrmuizel
This patch greatly improves the performance of QCMS transformations on x86 &
x86_64 systems.  Some notes:

0. On 32-bit x86 systems it does runtime selection between non-SIMD, SSE, and
SSE2 code paths.

1. On x86_64 systems the SSE2 code path is always taken.  The non-SIMD and SSE
code paths are left intact, but contemporary versions of the GCC and MSVC
compilers will see that they cannot be reached and optimize them away.

2. The execution of the SSE2 code path is reduced by 67%, relative to the
original Intel/Microsoft formatted ASM code.  The relative performance is seen
on a Pentium4 (Northwood) 2.4GHz CPU with DDR1 RAM.

3. The SSE code path provides a 80% reduction in execution time, relative to
the non-SIMD code path.  The relative performance is seen on a Pentium3
(Coppermine) 1.26GHz CPU with SDRAM.

4. The code has been split out into separate files so that it can be built
with different cflags (-msse, and -msse2) when using gcc.

5. Try to land again, this time with __attribute__((__force_align_arg_pointer__))
to avoid crashes on linux.
2009-10-01 14:19:41 -04:00
Jeff Muizelaar
fe18ffef29 Backed out changeset cb4f078cc8cb (bug 512865)
Was causing crashes on the leak test box.
2009-09-24 23:35:53 -04:00
Steve Snyder
0ed3ad35a2 Bug 512865. qcms: Improve SSE2 performance, add SSE support. r=jrmuizel
This patch greatly improves the performance of QCMS transformations on x86 &
x86_64 systems.  Some notes:

0. On 32-bit x86 systems it does runtime selection between non-SIMD, SSE, and
SSE2 code paths.

1. On x86_64 systems the SSE2 code path is always taken.  The non-SIMD and SSE
code paths are left intact, but contemporary versions of the GCC and MSVC
compilers will see that they cannot be reached and optimize them away.

2. The execution of the SSE2 code path is reduced by 67%, relative to the
original Intel/Microsoft formatted ASM code.  The relative performance is seen
on a Pentium4 (Northwood) 2.4GHz CPU with DDR1 RAM.

3. The SSE code path provides a 80% reduction in execution time, relative to
the non-SIMD code path.  The relative performance is seen on a Pentium3
(Coppermine) 1.26GHz CPU with SDRAM.

4. The code has been split out into separate files so that it can be built
with different cflags (-msse, and -msse2) when using gcc.
2009-09-24 11:58:24 -04:00
Peter Weilbacher
ad01d91539 Bug 488468: qcms build break with pre-4 GCCs (#if 0 commented text should be C/C++ commented), r=jmuizelaar 2009-04-23 19:38:11 +02:00
Jeff Muizelaar
eafab45f5f Bug 489133 - Handle bad cHRM chunks in PNGs. r=joe 2009-04-21 22:21:48 -04:00
Jeff Muizelaar
09f5d6e3c0 Bug 487566 - OS/2 problems to build qcms; r=mozilla@Weilbacher.org 2009-04-11 09:53:40 +02:00
Jeff Muizelaar
9980cb2188 Bug 481926 - Rewrite color management component sr=vlad, r=ted, r=joedrew (\o/)
Replaces lcms with qcms
2009-04-07 12:02:11 -04:00
Joe Drew
fc5c15825a Backed out changeset 6f3c2171bbb2:
Bug 481926 - Rewrite color management component. r=joe,ted sr=vlad
2009-04-03 16:27:27 -04:00
Jeff Muizelaar
eaef2088cb Bug 481926 - Rewrite color management component. r=joe,ted sr=vlad 2009-04-03 15:09:56 -04:00