Due to hysterical raisins, most RISC-V Linux distributions target a
RV64GC baseline excluding the Bit-manipulation ISA extensions, most
notably:
- Zba: address generation extension and
- Zbb: basic bit manipulation extension.
Most CPUs that would make sense to run FFmpeg on support Zba and Zbb
(including the current FATE runner), so it makes sense to optimise for
them. In fact a large chunk of existing assembler optimisations relies
on Zba and/or Zbb.
Since we cannot patch shared library code, the next best thing is to
carry a flag initialised at load-time and check it on need basis.
This results in 3 instructions overhead on isolated use, e.g.:
1: AUIPC rd, %pcrel_hi(ff_rv_zbb_supported)
LBU rd, %pcrel_lo(1b)(rd)
BEQZ rd, non_Zbb_fallback_code
// Zbb code here
The C compiler will typically load the flag ahead of time to reducing
latency, and can also keep it around if Zbb is used multiple times in a
single optimisation scope. For this to work, the flag symbol must be
hidden; otherwise the optimisation degrades with a GOT look-up to
support interposition:
1: AUIPC rd, GOT_OFFSET_HI
LD rd, GOT_OFFSET_LO(rd)
LBU rd, (rd)
BEQZ rd, non_Zbb_fallback_code
// Zbb code here
This patch adds code to provision the flag in libraries using bit
manipulation functions from libavutil: byte-swap, bit-weight and
counting leading or trailing zeroes.
Instead of an ad-hoc scheme. Also, combine skipping RASL frames with
skip_frame handling - current code seems flawed as it only executes for
the first slice of a RASL frame and unnecessarily unsets is_decoded,
which should not be set at this point anyway..
Some RASL frames in fate-hevc-afd-tc-sei that were previously discarded
are now output.
B0 is defined by system header, see f0f596dbc6 for ref.
Reviewed-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
src is apparently not guaranteed to be >8 byte aligned, but align to 16
nonetheless as the x86 asm will do unaligned loads anyway.
dst is guaranteed to be 32 byte aligned for the Y plane, but 16 byte for UV.
Signed-off-by: James Almer <jamrial@gmail.com>
The MMXEXT versions of the rgb2rgb functions tested here
always emit emms on their own. Therefore one can use
a stricter test to ensure that it stays that way.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
There is no benefit in using it: The fast path of copying
is not taken because of misalignment; furthermore we are
only dealing with a few byte here anyway, so simply copy
the bytes manually, avoiding the dependency on bitstream.c
in lavf (which also contains a function that is completely
unused in lavf).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
A 360 video specific tool
see https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9503377
passed files:
DMVR_A_Huawei_3.bit
WRAP_D_InterDigital_4.bit
WRAP_A_InterDigital_4.bit
WRAP_B_InterDigital_4.bit
WRAP_C_InterDigital_4.bit
ERP_A_MediaTek_3.bit
The OS may silently fix (emulate) unaligned hardware access exceptions.
This is extremely slow and code should be fixed not to rely on unaligned
access on affected hardware. Accordingly this requests that the OS
disable emulation and instead throw Bus error, which will be caught by
checkasm's signal handler.
This has no effects if the hardware supports unaligned access in
hardware, since no exceptions are generated. prctl() will fail safe in
that case.
The line width 8 is supposed to test corner case, while the
performance doesn't matter. Width 1080 is also a case of
unaligned to 16.
Width 1920 meant for benchmark (together with --runs options).
Signed-off-by: James Almer <jamrial@gmail.com>
From Benjamin Bross:
> for ALF where functions are in increments of 4 while 8 should be sufficient according to the spec.
Signed-off-by: Wu Jianhua <toqsxw@outlook.com>
According to the VVC specification (section 8.5.1), the maximum width/height of a subblock passed for DMVR SAD is 16. This along with previous constraint requiring width * height >= 128 means that 8x16, 16x8, and 16x16 are the only allowed sizes.
This changes check_vvc_sad() to only test and benchmark those sizes.
Several files already had standard license header (namely
2-clause BSD files), yet due to the 80 char line length limit,
they were not treated as such by source-check.sh (which
fate-source uses). Therefore relax the BSD check.
Reviewed-by: Rémi Denis-Courmont <remi@remlab.net>
Reviewed-by: Pierre-Anthony Lemieux <pal@sandflow.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
VVC does not have MMX code at all, so one can use the stricter
declare_func to also check that the MMX state has not been clobbered
with (which would be an ABI violation).
Reviewed-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The loop filters can write before the pointer given to them;
the actual test invocations correctly used an offset, while
the benchmark calls were lacking an offset. Therefore, when
running with benchmarking, these tests could have spurious
failures.
Signed-off-by: Martin Storsjö <martin@martin.st>
Some timers on certain device and test combinations can produce noisy
results, affecting the reliability of performance measurements. One
notable example of this is the Canaan K230 RISC-V development board.
An option to adjust the number of samples by an exponent (--runs) has
been added, allowing developers to increase the sample count for more
reliable results.
Signed-off-by: J. Dekker <jdek@itanimul.li>
Don't benchmark every single combination of widths and heights;
only benchmark cases which are squares (like in vvc_mc.c).
Contrary to vvc_mc, which increases sizes by doubling dimensions,
vvc_alf tests all sizes in increments of 4. Limit benchmarking to
the cases which are powers of two.
This reduces the number of benchmarked cases from 3072 down to 18.
This commit is the analog of 3f11eac757
for decoding: It sets the AV_FRAME_FLAG_KEY and (for video decoders)
also pict_type to AV_PICTURE_TYPE_I. It furthermore stops setting
audio frames as always being key frames -- it is wrong for e.g.
TrueHD/MLP. The latter also affects TAK and DFPWM.
The change already improves output for several decoders where
it has been forgotten to set e.g. pict_type like speedhq, wnv1
or tiff. The latter is the reason for the change to the exif-image-tiff
FATE test reference file.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Fixes "signed integer overflow: [varies] * 104858 cannot be represented in type 'int'" errors
under ubsan.
Signed-off-by: James Almer <jamrial@gmail.com>
