There is no MMX code for loop filters since commit
6a551f1405, so use declare_func
instead of declare_func_emms() to also test that we are not
in MMX mode after return.
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Unfortunately, it is common, and will remain so, that the Bit
manipulations are not enabled at compilation time. This is an official
policy for Debian ports in general (though they do not support RISC-V
officially as of yet) to stick to the minimal target baseline, which
does not include the B extension or even its Zbb subset.
For inline helpers (CPOP, REV8), compiler builtins (CTZ, CLZ) or
even plain C code (MIN, MAX, MINU, MAXU), run-time detection seems
impractical. But at least it can work for the byte-swap DSP functions.
RVV defines a total of 12 different extensions, including:
- 5 different instruction subsets:
- Zve32x: 8-, 16- and 32-bit integers,
- Zve32f: Zve32x plus single precision floats,
- Zve64x: Zve32x plus 64-bit integers,
- Zve64f: Zve32f plus Zve64x,
- Zve64d: Zve64f plus double precision floats.
- 6 different vector lengths:
- Zvl32b (embedded only),
- Zvl64b (embedded only),
- Zvl128b,
- Zvl256b,
- Zvl512b,
- Zvl1024b,
- and the V extension proper: equivalent to Zve64f and Zvl128b.
In total, there are 6 different possible sets of supported instructions
(including the empty set), but for convenience we allocate one bit for
each type sets: up-to-32-bit ints (RVV_I32), floats (RVV_F32),
64-bit ints (RVV_I64) and doubles (RVV_F64).
Whence the vector size is needed, it can be retrieved by reading the
unprivileged read-only vlenb CSR. This should probably be a separate
helper macro if needed at a later point.
This introduces compile-time and run-time CPU detection on RISC-V. In
practice, I doubt that FFmpeg will ever see a RISC-V CPU without all of
I, F and D extensions, and if it does, it probably won't have run-time
detection. So the flags are essentially always set.
But as things stand, checkasm wants them that way. Compare the ARMV8
flag on AArch64. We are nowhere near running short on CPU flag bits.
~4x faster than the C version.
The shuffles in the 15pt dim1 are seriously expensive. Not happy with it,
but I'm contempt.
Can be easily converted to pure AVX by removing all vpermpd/vpermps
instructions.
Old one was written with the assumption only even inputs would be given.
This very messy replacement supports even and odd inputs, and supports
AVX2 for extra speed. The buffers given are usually quite big (4k samples),
so the speedup is worth it.
The new SSE version is still faster than the old inline asm version by 33%.
Also checkasm is provided to make sure this monstrosity works.
This fixes some FATE tests.
This avoids triggering overflows in the filters, and avoids stray
test failures in the approximate functions on x86; due to rounding
differences, one implementation might overflow while another one
doesn't.
Signed-off-by: Martin Storsjö <martin@martin.st>
Previously, the checkasm test always passed h=8, so no other cases
were tested.
Out of the me_cmp functions, in practice, some functions are hardcoded
to always assume a 8x8 block (ignoring the h parameter), while others
do use the parameter. For those with hardcoded height, both the
reference C function and the assembly implementations ignore the
parameter similarly.
The documentation for the functions indicate that heights between
w/2 and 2*w, within the range of 4 to 16, should be supported. This
patch just tests random heights in that range, without knowing what
width the current function actually uses.
Signed-off-by: Martin Storsjö <martin@martin.st>
Change the reference to exactly match the C reference in swscale,
instead of exactly matching the x86 SIMD implementations (which
differs slightly). Test with and without SWS_ACCURATE_RND - if this
flag isn't set, the output must match the C reference exactly,
otherwise it is allowed to be off by 2.
Mark a couple x86 functions as unavailable when SWS_ACCURATE_RND
is set - apparently this discrepancy hasn't been noticed in other
exact tests before.
Add a test for yuv2plane1.
Signed-off-by: Jonathan Swinney <jswinney@amazon.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
This codepath is enabled by default on arm, if the linux perf API
is available, unless disabled with --disable-linux-perf.
Signed-off-by: Martin Storsjö <martin@martin.st>
- ff_pix_abs16_neon
- ff_pix_abs16_xy2_neon
In direct micro benchmarks of these ff functions verses their C implementations,
these functions performed as follows on AWS Graviton 3.
ff_pix_abs16_neon:
pix_abs_0_0_c: 141.1
pix_abs_0_0_neon: 19.6
ff_pix_abs16_xy2_neon:
pix_abs_0_3_c: 269.1
pix_abs_0_3_neon: 39.3
Tested with:
./tests/checkasm/checkasm --test=motion --bench --disable-linux-perf
Signed-off-by: Jonathan Swinney <jswinney@amazon.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
SSE3 instruction movdqa in ff_yuv2yuvX_sse3() expects a 16-byte aligned address for a memory address, or else a segfault is generated.
The src_pixels buffer below was not aligned to 16 bytes on the stack necessarily, so we got segfaults during fate-checkasm-sw_scale.
Therefore 16-byte align all of these local variables, aligning them too much shouldn't hurt.
The HEVC decoder can call these functions with smaller widths than the
functions themselves are designed to operate on so we should only check
the relevant output
Signed-off-by: J. Dekker <jdek@itanimul.li>
This removes a dependency of checkasm on lavc/v210_enc.o
and also allows to inline ff_v210enc_init() irrespectively of
interposing.
This dependency pulled basically all of libavcodec into checkasm,
in particular all codecs.
This also makes checkasm work when using shared Windows builds:
On Windows, it needs to be known to the compiler whether a data
symbol is external to the library/executable or not; hence the
need for av_export_avutil. checkasm needs access to the internals
of the libraries it tests and is therefore linked statically to all
the libraries. This means that the users of avpriv_cga_font and
avpriv_vga16_font in libavcodec (namely ansi.o, bintext.o, tmv.o)
end up in the same executable as the symbols, although they have
been compiled as if these symbols were external, leading to linker
errors. With this commit said files are discarded by the linker,
bypassing this problem.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This removes a dependency of checkasm on lavc/v210_dec.o
and also allows to inline ff_v210dec_init() irrespectively of
interposing.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This removes a dependency of checkasm on lavfi/vf_threshold.o
and also allows to inline ff_threshold_init() irrespectively of
interposing.
With this patch checkasm no longer pulls all of lavfi and lavf in.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This removes a dependency of checkasm on lavfi/vf_nlmeans.o
and also allows to inline ff_nlmeans_init() irrespectively of
interposing.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This removes a dependency of checkasm on lavfi/vf_hflip.o
and also allows to inline ff_hflip_init() irrespectively of
interposing.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This removes a dependency of checkasm on lavfi/vf_gblur.o
and also allows to inline ff_gblur_init() irrespectively of
interposing.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This removes a dependency of checkasm on lavfi/vf_blend.o
and also allows to inline ff_blend_init() irrespectively of
interposing.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Only the AudioFIRDSPContext and the functions for its initialization
are needed outside of lavfi/af_afir.c.
Also rename the header to af_afirdsp.h to reflect the change.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This test deliberately doesn't exercise the full range of inputs described in
the committee draft VC-1 standard. It says:
input coefficients in frequency domain, D, satisfy -2048 <= D < 2047
intermediate coefficients, E, satisfy -4096 <= E < 4095
fully inverse-transformed coefficients, R, satisfy -512 <= R < 511
For one thing, the inequalities look odd. Did they mean them to go the
other way round? That would make more sense because the equations generally
both add and subtract coefficients multiplied by constants, including powers
of 2. Requiring the most-negative values to be valid extends the number of
bits to represent the intermediate values just for the sake of that one case!
For another thing, the extreme values don't look to occur in real streams -
both in my experience and supported by the following comment in the AArch32
decoder:
tNhalf is half of the value of tN (as described in vc1_inv_trans_8x8_c).
This is done because sometimes files have input that causes tN + tM to
overflow. To avoid this overflow, we compute tNhalf, then compute
tNhalf + tM (which doesn't overflow), and then we use vhadd to compute
(tNhalf + (tNhalf + tM)) >> 1 which does not overflow because it is
one instruction.
My AArch64 decoder goes further than this. It calculates tNhalf and tM
then does an SRA (essentially a fused halve and add) to compute
(tN + tM) >> 1 without ever having to hold (tNhalf + tM) in a 16-bit element
without overflowing. It only encounters difficulties if either tNhalf or
tM overflow in isolation.
I haven't had sight of the final standard, so it's possible that these
issues were dealt with during finalisation, which could explain the lack
of usage of extreme inputs in real streams. Or a preponderance of decoders
that only support 16-bit intermediate values in their inverse transforms
might have caused encoders to steer clear of such cases.
I have effectively followed this approach in the test, and limited the
scale of the coefficients sufficient that both the existing AArch32 decoder
and my new AArch64 decoder both pass.
Signed-off-by: Ben Avison <bavison@riscosopen.org>
Signed-off-by: Martin Storsjö <martin@martin.st>
Note that the benchmarking results for these functions are highly dependent
upon the input data. Therefore, each function is benchmarked twice,
corresponding to the best and worst case complexity of the reference C
implementation. The performance of a real stream decode will fall somewhere
between these two extremes.
Signed-off-by: Ben Avison <bavison@riscosopen.org>
Signed-off-by: Martin Storsjö <martin@martin.st>
This avoids unnecessary rebuilds of most source files if only the
list of enabled components has changed, but not the other properties
of the build, set in config.h.
Signed-off-by: Martin Storsjö <martin@martin.st>
