Christophe Gisquet 996697e266 x86: float dsp: unroll SSE versions ...

vector_fmul and vector_fmac_scalar are guaranteed that they can process in
batch of 16 elements, but their SSE versions only does 8 at a time.

Therefore, unroll them a bit.
299 to 261c for 256 elements in vector_fmac_scalar on Arrandale/Win64.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>

2014-02-20 14:18:05 +01:00

..

asm.h

dsputil: Make dsputil selectable

2013-04-10 11:04:05 +03:00

bswap.h

x86: place some inline asm under #if HAVE_INLINE_ASM

2012-06-25 13:23:12 +01:00

cpu.c

libavutil: x86: Add AVX2 capable CPU detection.

2013-10-25 19:36:55 +01:00

cpu.h

libavutil: x86: Add AVX2 capable CPU detection.

2013-10-25 19:36:55 +01:00

cpuid.asm

x86: include x86inc.asm in x86util.asm

2012-10-31 00:37:42 +01:00

emms.asm

x86: Add a Yasm-based emms() replacement

2013-01-18 22:02:13 +01:00

emms.h

avutil: Ensure that emms_c is always defined, even on non-x86

2013-02-14 19:29:04 +01:00

float_dsp_init.c

Consistently use "cpu_flags" as variable/parameter name for CPU flags

2013-07-18 00:31:35 +02:00

float_dsp.asm

x86: float dsp: unroll SSE versions

2014-02-20 14:18:05 +01:00

intreadwrite.h

Replace FFmpeg with Libav in licence headers

2011-03-19 13:33:20 +00:00

lls_init.c

x86: lpc: simd av_evaluate_lls

2013-06-29 13:23:57 +02:00

lls.asm

lls/x86: use 3-operator vaddpd in ADDPD_MEM

2013-07-02 10:15:09 +02:00

Makefile

x86: lpc: simd av_update_lls

2013-06-29 13:23:57 +02:00

timer.h

avutil: Fix compilation with inline asm disabled on mingw

2013-09-22 00:50:32 +03:00

w64xmmtest.h

Add more missing includes after removing the implicit common.h

2012-08-16 10:49:54 +03:00

x86inc.asm

x86inc: Speed up assembling with Yasm

2014-01-26 18:40:08 +01:00

x86util.asm

x86inc: FMA3/4 Support

2013-10-14 12:41:54 +01:00