Replace INIT_AVX by INIT_XMM avx. Port the whole file to use cpuflag
based function declarations. Remove (now unused) cputype argument in
function declaration macros. Change function prototypes to have mmx2
instead of mmxext as suffix, since that's required by cpuflags.
Instead of using an evil VLA, fall back to C version when edge
emulation is needed. MPEG4 GMC is a rarely used fringe feature
so the speed loss is an acceptable cost for safer code.
Signed-off-by: Mans Rullgard <mans@mansr.com>
For some reason add_hfyu_median_prediction_cmov is only selected
on 3Dnow-capable CPUs, even though it uses no 3Dnow instructions.
This patch allows it to be selected on any cpu with cmov with the
possibility of being overridden by the mmxext version.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The optimized dct_quantize template functions reference optimized
fdct symbols, so these functions must only be enabled if the relevant
optimizations have been enabled by configure.
This fixes two issues preventing suncc from building this code.
The undocumented 'a' operand modifier, causing gcc to omit a $ in
front of immediate operands (as required in addresses), is not
supported by suncc. Luckily, the also undocumented 'c' modifer
has the same effect and is supported.
On some asm statements with a large number of operands, suncc for no
obvious reason fails to correctly substitute some of the operands.
Fortunately, some of the operands in these statements are plain
numbers which can be inserted directly into the code block instead
of passed as operands.
With these changes, the code builds correctly with both gcc and
suncc.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This code contains a C array of addresses of labels defined in
inline asm. To do this, the names must be declared as external
in C. The declared type does not matter since only the address is
used, and for some reason, the author of the code used the 'void'
type despite taking the address of a void expression being invalid.
Changing the type to char, a reasonable choice since the alignment
of the code labels cannot be known or guaranteed, eliminates gcc
warnings and allows building with suncc.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This macro is only used in two places, both in libavcodec, so this
is a more sensible place for it.
Two small tweaks to the macro are made:
- removing the trailing semicolon
- dropping unnecessary 'volatile' from the x86 asm
Signed-off-by: Mans Rullgard <mans@mansr.com>
yasm tolerates mismatch between movd/movq and source register size,
adjusting the instruction according to the register. nasm is more
strict.
Signed-off-by: Mans Rullgard <mans@mansr.com>
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.
This makes add_hfyu_left_prediction_sse4() handle sources that are not
16-byte aligned in its own function rather than by proxying the call to
add_hfyu_left_prediction_ssse3(). This fixes a crash on Win64, since the
sse4 version clobberes xmm6, but the ssse3 version (which uses MMX regs)
does not restore it, thus leading to XMM clobbering and RSP being off.
Fixes bug 342.
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.
Some calculations were changed in b6a3849 to use mmsize, which was not correct
for the AVX version, which uses INIT_YMM and therefore has mmsize == 32.
Fixes Bug 341.
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
These functions are not faster than other mmx implementations on
any hardware I have been able to test on, and they are horribly
inaccurate. There is thus no reason to ever use them.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This completes the conversion of h264dsp to yasm; note that h264 also
uses some dsputil functions, most notably qpel. Performance-wise, the
yasm-version is ~10 cycles faster (182->172) on x86-64, and ~8 cycles
faster (201->193) on x86-32.
Without this, cglobal will expand "z" to "zh" to access the high byte
in a register's word, which causes a name collision with the ZH(x) macro
further up in this file.
Mixing yasm and inline asm is a bad idea, since if either yasm or inline
asm is not supported by your toolchain, all of the asm stops working.
Thus, better to use either one or the other alone.
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
In ff_put_pixels_clamped_mmx(), there are two assembly code blocks.
In the first block (in the unrolled loop), the instructions
"movq 8%3, %%mm1 \n\t", and so forth, have problems.
From above instruction, it is clear what the programmer wants: a load from
p + 8. But this assembly code doesn’t guarantee that. It only works if the
compiler puts p in a register to produce an instruction like this:
"movq 8(%edi), %mm1". During compiler optimization, it is possible that the
compiler will be able to constant propagate into p. Suppose p = &x[10000].
Then operand 3 can become 10000(%edi), where %edi holds &x. And the instruction
becomes "movq 810000(%edx)". That is, it will stride by 810000 instead of 8.
This will cause a segmentation fault.
This error was fixed in the second block of the assembly code, but not in
the unrolled loop.
How to reproduce:
This error is exposed when we build using Intel C++ Compiler, with
IPO+PGO optimization enabled. Crashed when decoding an MJPEG video.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
This moves all VP3-specific function pointers from dsputil to a
new vp3dsp context. There is no reason to ever use the VP3 IDCT
where an MPEG2 IDCT is expected or vice versa.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The SPLATB_REG macro already adds the 'd' suffix internally.
This fixes building on Win64, which has been broken since 878e66902.
This worked for unix, where r2 happened to be rdx in this case, which
with the first suffix rdxd was mapped to eax, and eaxd is defined back
to eax. On win64 however, r2 happened to be R8 in this case, and
R8d mapps to R8D just fine, but there's no mapping for R8Dd to anything.
Signed-off-by: Martin Storsjö <martin@martin.st>
The problem is that the ssse3 psign instruction does the wrong
thing here. Commit ea60dfe incorrectly removed a macro emulating
this instruction for pre-ssse3 code. However, the emulation is
incorrect, and the code relies on the behaviour of the macro.
Specifically, the psign sets destination elements to zero where
the corresponding source element is zero, whereas the emulation
only negates destination elements where the source is negative.
Furthermore, the PSIGNW_MMX macro in x86util.asm is totally bogus,
which is why the original VC-1 code had an additional right shift
when using it. Since the psign instruction cannot be used here,
skip all the macro hell and use the working instruction sequence
directly.
None of this was noticed due a stray return statement in
ff_vc1dsp_init_mmx() which meant that only the mmx version of the
loop filter was ever used (before being removed in ea60dfe).
Signed-off-by: Mans Rullgard <mans@mansr.com>
The function call was a mess to handle, and memcpy cannot make
the assumptions we do in the new code.
Tested on an IMC sample: 430c -> 370c.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This removes a dependency on implementation details from generic
code and allows easy addition of the equivalent optimisation for
other architectures than x86.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Code mostly inspired by vp8's MC, however:
- its MMX2 horizontal filter is worse because it can't take advantage of
the coefficient redundancy
- that same coefficient redundancy allows better code for non-SSSE3 versions
Benchmark (rounded to tens of unit):
V8x8 H8x8 2D8x8 V16x16 H16x16 2D16x16
C 445 358 985 1785 1559 3280
MMX* 219 271 478 714 929 1443
SSE2 131 158 294 425 515 892
SSSE3 120 122 248 387 390 763
End result is overall around a 15% speedup for SSSE3 version (on 6 sequences);
all loop filter functions now take around 55% of decoding time, while luma MC
dsp functions are around 6%, chroma ones are 1.3% and biweight around 2.3%.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
This adds a hand-optimized assembly version for get_cabac much like the
existing one, but it works if the table offsets are RIP-relative.
Compared to the non-RIP-relative version this adds 2 lea instructions
and it needs one extra register. get_cabac() gets about 40% faster, for
an overall speedup of about 5%.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
The reason is this is easier for PIC code (in particular on darwin...).
Keep the old names as pointers (static in cabac_functions.h so gcc
knows these are just immediate offsets) so the c code can nicely stay the same
(alternatively could use offsets directly in the functions needing the
tables). This should produce the same code as before with non-pic and better
code (confirmed) with pic.
The assembly uses the new table but still won't work for PIC case.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
This removes all references to AVCodecContext.dsp_mask and marks
it for eviction at the next version bump. It has been superseded
by av_set_cpu_flag_mask() which, unlike this field, works everywhere.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Recent register allocation changes (x86inc.asm update) changed the
register order and thus opcodes for the inner loops. One of them became
>128bytes, which confuses other parts of this function where it jumps
to fixed-offset positions to extend the edge by fixed amounts. A simple
register change fixes this.
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>
Quite often, the original weights are multiple of 512. By prescaling them
by 1/512 when they are computed (once per frame), no intermediate shifting
is needed, and no prescaling on each call either.
The x86 code already used that trick.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
Prevents a signflip in the counter, and a subsequent crash because of
overreads/overwrites.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC: libav-stable@libav.org
Since the values are floats, using the float operations
makes sense, improves performance on some CPUs and
makes the code SSE compatible instead of needing SSE2.
Based on suggestion by Jason.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
There is only one caller, which does not need the shifting. Other use cases
are situations where different roundings would be needed.
The x86 and neon versions are modified accordingly.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
movq from SSE register _to_ memory is an SSE2 instruction.
Use the SSE movlps function instead that does the same thing.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
This splits ff_dsputil_init_mmx() into multiple functions, one for
each MMX/SSE level, somewhat simplifying the nested conditions.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Unrolling the main loop to process, instead of 4 elements:
- 8: minor gain of 2 cycles (not worth the extra object size)
- 2: loss of 8 cycles.
Assigning STEP to a register is a loss. Output address (Y) is almost always
unaligned.
Timings:
- C (32/64 bits): 117/109 cycles
- SSE: 57 cycles
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
The 32bits targets have been compiled with -mfpmath=sse for proper reference.
sbr_sum_square C /32bits: 82c (unrolled)/102c
C /64bits: 69c (unrolled)/82c
SSE/32bits: 42c
SSE/64bits: 31c
Use of SSE4.1 dpps to perform the final sum is slower.
Not unrolling to perform 8 operations in a loop yields 10 more cycles.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
This prevents having to sign-extend on 64-bit systems with 32-bit ints,
such as x86-64. Also fixes crashes on systems where we don't do it and
arguments are not in registers, such as Win64 for all weight functions.
By replacing memcpy with an unrolled loop using the alignment knowledge
it has, some speedup can be obtained.
Before (gcc 4.6.1): ~400 cycles
After: ~370 cycles
Overall, around 2% speed increase when decoding a 2400s mp3 to f32le.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
We need to do unsigned saturation in order to cover the corner case when the
absolute coefficient value is 16777215 (the maximum value).
Fixes Bug #216
This will be useful to test more aggressively for failures to mark XMM
registers as clobbered in Win64 builds, and prevent regressions thereof.
Based on a patch by Ramiro Polla <ramiro.polla@gmail.com>
Provide MMX, SSE2 and SSSE3 versions, with a fast-path when the weights are
multiples of 512 (which is often the case when the values round up nicely).
*_TIMER report for the 16x16 and 8x8 cases:
C:
9015 decicycles in 16, 524257 runs, 31 skips
2656 decicycles in 8, 524271 runs, 17 skips
MMX:
4156 decicycles in 16, 262090 runs, 54 skips
1206 decicycles in 8, 262131 runs, 13 skips
MMX on fast-path:
2760 decicycles in 16, 524222 runs, 66 skips
995 decicycles in 8, 524252 runs, 36 skips
SSE2:
2163 decicycles in 16, 262131 runs, 13 skips
832 decicycles in 8, 262137 runs, 7 skips
SSE2 with fast path:
1783 decicycles in 16, 524276 runs, 12 skips
711 decicycles in 8, 524283 runs, 5 skips
SSSE3:
2117 decicycles in 16, 262136 runs, 8 skips
814 decicycles in 8, 262143 runs, 1 skips
SSSE3 with fast path:
1315 decicycles in 16, 524285 runs, 3 skips
578 decicycles in 8, 524286 runs, 2 skips
This means around a 4% speedup for some sequences.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
While pshufb allows emulating bswap on XMM registers for SSSE3, more
shuffling is needed for SSE2. Alignment is critical, so specific codepaths
are provided for this case.
For the huffyuv sequence "angels_480-huffyuvcompress.avi":
C (using bswap instruction): ~ 55k cycles
SSE2: ~ 40k cycles
SSSE3 using unaligned loads: ~ 35k cycles
SSSE3 using aligned loads: ~ 30k cycles
Signed-off-by: Diego Biurrun <diego@biurrun.de>
