Commit Graph

88 Commits

Author SHA1 Message Date
Diego Biurrun
a7329e5fc2 x86: get_cpu_flags: add necessary ifdefs around function body
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file.  Surrounding the function body with ifdefs allows
building even when cpuid is not defined.  An empty cpuflags mask is
returned in this case.
2012-10-04 19:29:14 +02:00
Diego Biurrun
f6fbce761e x86: Drop CPU detection intrinsics
Now that there is CPU detection in YASM, there will always be one of
inline or external assembly enabled, which obviates the need to fall
back on CPU detection through compiler intrinsics.
2012-10-04 19:29:14 +02:00
Diego Biurrun
1f6d86991f x86: Add YASM implementations of cpuid and xgetbv from x264
This allows detecting CPU features with builds that have neither
gcc inline assembly nor the right compiler intrinsics enabled.
2012-10-04 19:29:14 +02:00
Diego Biurrun
54b243141e x86: cpu: Break out test for cpuid capabilities into separate function 2012-10-04 18:09:21 +02:00
Diego Biurrun
cc5e9e5ff0 x86: ff_get_cpu_flags_x86(): Avoid a pointless variable indirection 2012-10-04 17:58:42 +02:00
Diego Biurrun
e0c6cce447 x86: Replace checks for CPU extensions and flags by convenience macros
This separates code relying on inline from that relying on external
assembly and fixes instances where the coalesced check was incorrect.
2012-09-08 18:18:34 +02:00
Justin Ruggles
7327525997 x86: float_dsp: fix ff_vector_fmac_scalar_avx() on Win64
The SWAP macro does not work for explicit xmm/ymm usage, so instead just move
the scalar value from xmm2 to xmm0.
2012-09-07 14:49:10 -04:00
Diego Biurrun
f82c4fb27f x86: Add convenience macros to check for CPU extensions and flags 2012-09-04 01:44:59 +02:00
Diego Biurrun
17337f54c0 x86: Split inline and external assembly #ifdefs 2012-08-31 01:53:25 +02:00
Diego Biurrun
a886b279a0 x86: cosmetics: Comment some #endifs for better readability 2012-08-30 18:50:33 +02:00
Loren Merritt
7a1944b907 vf_hqdn3d: x86 asm
13% faster on penryn, 16% on sandybridge, 15% on bulldozer
Not simd; a compiler should have generated this, but gcc didn't.
2012-08-26 10:49:14 +00:00
Justin Ruggles
6092dafb5a lavr: x86: optimized 6-channel s16 to fltp conversion 2012-08-23 20:10:57 -04:00
Mans Rullgard
5b170c0bea x86: remove FASTDIV inline asm
GCC 4.3 and later do the right thing with the plain C code.  Earlier
versions in 32-bit mode generate one extra instruction, needlessly
zeroing what would be the high half of the shifted value.  At least
two gcc configurations miscompile the inline asm in some situations.

In 64-bit mode, all gcc versions generate imul r64, r64 followed by
shr.  On Intel i7 and later, this imul is faster 32-bit mul.  On
older Intel and all AMD, it is slightly slower.  On Atom it is much
slower.

Considering where the FASTDIV macro is used, any overall negative
performance impact of this change should be negligible.  If anyone
cares, they should file a bug against gcc and get the instruction
selection fixed.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-08-22 14:29:10 +01:00
Martin Storsjö
33e112847d Add more missing includes after removing the implicit common.h
Signed-off-by: Martin Storsjö <martin@martin.st>
2012-08-16 10:49:54 +03:00
Martin Storsjö
70766c2182 Add some more missing includes after removing the implicit common.h
Signed-off-by: Martin Storsjö <martin@martin.st>
2012-08-15 23:48:48 +03:00
Mans Rullgard
070a402b60 x86: move MANGLE() and related macros to libavutil/x86/asm.h
These x86-specific macros do not belong in generic code.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-08-09 00:58:20 +01:00
Mans Rullgard
c318626ce2 x86: rename libavutil/x86_cpu.h to libavutil/x86/asm.h
This puts x86-specific things in the x86/ subdirectory where they
belong.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-08-09 00:58:20 +01:00
Mans Rullgard
edd8226795 x86: fix build with nasm 2.08
It appears that something goes wrong in old nasm versions when the
%+ operator is used in the last argument of a macro invocation and
this argument is tested with %ifdef within the macro.  This patch
rearranges the macro arguments such that the %+ operator is never
used in the last argument.
2012-08-07 15:24:34 +01:00
Mans Rullgard
180d43bc67 x86: use nop cpu directives only if supported
nasm does not support 'CPU foonop' directives.  This adds a configure
test for the directive and uses it only if supported.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-08-07 15:22:20 +01:00
Mans Rullgard
7238265052 x86: fix rNmp macros with nasm
For some reason, nasm requires this.  No harm done to yasm.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-08-07 15:21:58 +01:00
Mans Rullgard
a3df4781f4 x86: add colons after labels
nasm prints a warning if the colon is missing.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-08-07 15:20:56 +01:00
Diego Biurrun
239fdf1b4a x86: build: replace mmx2 by mmxext
Refactoring mmx2/mmxext YASM code with cpuflags will force renames.
So switching to a consistent naming scheme beforehand is sensible.
The name "mmxext" is more official and widespread and also the name
of the CPU flag, as reported e.g. by the Linux kernel.
2012-08-03 22:51:05 +02:00
Diego Biurrun
ca844b7be9 x86: Use consistent 3dnowext function and macro name suffixes
Currently there is a wild mix of 3dn2/3dnow2/3dnowext.  Switching to
"3dnowext", which is a more common name of the CPU flag, as reported
e.g. by the Linux kernel, unifies this.
2012-08-03 14:00:47 +02:00
Loren Merritt
f8d8fe255d x86inc: clip num_args to 7 on x86-32.
This allows us to unconditionally set the cglobal num_args
parameter to a bigger value, thus making writing yasm code
even easier than before.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
2012-07-28 08:29:45 -07:00
Ronald S. Bultje
96c9cc1094 x86inc: sync to latest version from x264. 2012-07-28 08:29:44 -07:00
Justin Ruggles
79687079a9 x86: add support for fmaddps fma4 instruction with abstraction to avx/sse 2012-07-27 11:25:48 -04:00
Ronald S. Bultje
30b45d9c38 x86inc: automatically insert vzeroupper for YMM functions. 2012-07-26 13:43:16 -07:00
Jason Garrett-Glaser
85a3c19ed1 dsputil: x86: add SHUFFLE_MASK_W macro
Simplifies pshufb masks that operate on words.
2012-07-22 16:56:58 -04:00
Ronald S. Bultje
358d854df8 x86/cpu: implement get/set_eflags using intrinsics
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Signed-off-by: Martin Storsjö <martin@martin.st>
2012-07-10 14:33:32 +03:00
Ronald S. Bultje
c0ee695bd7 x86/cpu: implement support for cpuid through intrinsics
Signed-off-by: Martin Storsjö <martin@martin.st>
2012-07-10 14:33:24 +03:00
Ronald S. Bultje
3f150ffba3 x86/cpu: implement support for xgetbv through intrinsics
Signed-off-by: Martin Storsjö <martin@martin.st>
2012-07-10 14:33:17 +03:00
Ronald S. Bultje
07b287020c x86/timer: implement an intrinsic-based version for rdtsc (AV_READ_TIME). 2012-07-07 13:35:07 -07:00
Loren Merritt
4d4752366f x86inc: add SPLATB_LOAD, SPLATB_REG, PSHUFLW macros
Signed-off-by: Diego Biurrun <diego@biurrun.de>
2012-07-05 17:37:11 +02:00
Loren Merritt
2cd1f5cadc x86inc: modify ALIGN to not generate long nops on i586
Signed-off-by: Diego Biurrun <diego@biurrun.de>
2012-07-05 17:37:11 +02:00
Mans Rullgard
889c1ec4cc x86: cpu: clean up check for cpuid instruction support
This adds macros for accessing the EFLAGS register and uses
these instead of coding the entire check in inline asm.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-07-01 12:25:33 +01:00
Mans Rullgard
963cdf39b4 x86: cpu: whitespace (mostly) cosmetics
This adds whitespace around operators, aligns line continuation
backslashes, and breaks long lines.  Also fixes an ifdef halfway
through a statement.  The one line of duplication this saved is
not worth the ugliness.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-06-25 16:24:31 +01:00
Ronald S. Bultje
8123e0901f x86: place some inline asm under #if HAVE_INLINE_ASM
Signed-off-by: Mans Rullgard <mans@mansr.com>
2012-06-25 13:23:12 +01:00
Diego Biurrun
65345a5a30 x86: Add CPU flag for the i686 cmov instruction 2012-06-23 16:21:50 +02:00
Justin Ruggles
82b2df9790 float_dsp: add x86-optimized functions for vector_fmac_scalar() 2012-06-18 18:01:14 -04:00
Justin Ruggles
d5a7229ba4 Add a float DSP framework to libavutil
Move vector_fmul() from DSPContext to AVFloatDSPContext.
2012-06-08 13:14:38 -04:00
Vitor Sessak
4a301706fd x86: Avoid movs on BUTTERFLYPS when in AVX mode
Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2012-05-29 15:29:46 +02:00
Justin Ruggles
5cc6d5244d lavr: replace the SSE version of ff_conv_fltp_to_flt_6ch() with SSE4 and AVX
The current SSE version is slower than the MMX version on Athlon64 and Sandy
Bridge, but the SSE4 and AVX versions are faster on Sandy Bridge.
2012-05-09 16:17:59 -04:00
Justin Ruggles
c8af852b97 Add libavresample
This is a new library for audio sample format, channel layout, and sample rate
conversion.
2012-04-24 21:28:27 -04:00
Loren Merritt
705f3d4759 x86inc: support AVX abstraction for 2-operand instructions
Add cvtdq2ps and cvtps2dq to the AVX instruction list.

Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
2012-04-18 21:14:32 -04:00
Diego Biurrun
baaab6069a build: Move all arch OBJS declarations into arch subdirectory Makefiles. 2012-04-12 21:30:13 +02:00
Henrik Gramner
729f90e268 x86inc improvements for 64-bit
Add support for all x86-64 registers
Prefer caller-saved register over callee-saved on WIN64
Support up to 15 function arguments

Also (by Ronald S. Bultje)
Fix up our asm to work with new x86inc.asm.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
2012-04-11 15:47:00 -04:00
Ronald S. Bultje
98b9da2ac7 x86inc: add *mp named argument support to DEFINE_ARGS. 2012-03-14 20:09:53 -07:00
Loren Merritt
0f53d0cf4b x86inc: don't "bake" stack_offset in named arguments.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
2012-03-03 20:39:59 -08:00
Haruhiko Yamagata
166f399377 x86inc: support yasm -f win64 flag also.
This sets __OUTPUT_FORMAT__ to win64 instead of win32, even though both
(through -m amd64) produce 64-bit binary code.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
2012-02-08 10:31:14 -08:00
Henrik Gramner
9cf7385309 x86inc: allow manual use of WIN64_SPILL_XMM.
Functions using INIT_MMX may still access XMM registers through direct
means (xmm0-15). Therefore, they still need to be marked for clobber
so they can be properly saved/restored.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
2012-02-08 10:31:14 -08:00