This commit adds new code paths for vscale when filterSize is 2, 4, or
8. By using specialized code with unrolling to match the filterSize we
can improve performance.
On AWS c7g (Graviton 3, Neoverse V1) instances:
before after
yuv2yuvX_2_0_512_accurate_neon: 558.8 268.9
yuv2yuvX_4_0_512_accurate_neon: 637.5 434.9
yuv2yuvX_8_0_512_accurate_neon: 1144.8 806.2
yuv2yuvX_16_0_512_accurate_neon: 2080.5 1853.7
Signed-off-by: Jonathan Swinney <jswinney@amazon.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Use scalar times vector multiply accumlate instructions instead of
vector times vector to remove the need for replicating load instructions
which are slightly slower.
On AWS c7g (Graviton 3, Neoverse V1) instances:
yuv2yuvX_8_0_512_accurate_neon: 1144.8 987.4
yuv2yuvX_16_0_512_accurate_neon: 2080.5 1869.4
Signed-off-by: Jonathan Swinney <jswinney@amazon.com>
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 specialization handles the case where filtersize is 4 mod 8, e.g.
12, 20, etc. Aarch64 was previously using the c function for this case.
This implementation speeds up that case significantly.
hscale_8_to_15__fs_12_dstW_512_c: 6234.1
hscale_8_to_15__fs_12_dstW_512_neon: 1505.6
Signed-off-by: Jonathan Swinney <jswinney@amazon.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
x64 always has MMX, MMXEXT, SSE and SSE2 and this means
that some functions for MMX, MMXEXT, SSE and 3dnow are always
overridden by other functions (unless one e.g. explicitly
disables SSE2). So given that the only systems that
benefit from these functions are truely ancient 32bit x86s
they are removed.
Moreover, some of the removed code was buggy/not bitexact
and lead to failures involving the f32le and f32be versions of
gray, gbrp and gbrap on x86-32 when SSE2 was not disabled.
See e.g.
https://fate.ffmpeg.org/report.cgi?time=20220609221253&slot=x86_32-debian-kfreebsd-gcc-4.4-cpuflags-mmx
Notice that yuv2yuvX_mmx is not removed, because it is used
by SSE3 and AVX2 as fallback in case of unaligned data and
also for tail processing. I don't know why yuv2yuvX_mmxext
isn't being used for this; an earlier version [1] of
554c2bc708 used it, but
the version that was eventually applied does not.
[1]: https://ffmpeg.org/pipermail/ffmpeg-devel/2020-November/272124.html
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
x64 always has MMX, MMXEXT, SSE and SSE2 and this means
that some functions for MMX, MMXEXT and 3dnow are always
overridden by other functions (unless one e.g. explicitly
disables SSE2) for x64. So given that the only systems that
benefit from these functions are truely ancient 32bit x86s
they are removed.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
x64 always has MMX, MMXEXT, SSE and SSE2 and this means
that some functions for MMX, MMXEXT and 3dnow are always
overridden by other functions (unless one e.g. explicitly
disables SSE2) for x64. So given that the only systems that
benefit from these functions are truely ancient 32bit x86s
they are removed.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is more spec-compliant because it does not rely
on dead-code elimination by the compiler. Especially
MSVC has problems with this, as can be seen in
https://ffmpeg.org/pipermail/ffmpeg-devel/2022-May/296373.html
or
https://ffmpeg.org/pipermail/ffmpeg-devel/2022-May/297022.html
This commit does not eliminate every instance where we rely
on dead code elimination: It only tackles branching to
the initialization of arch-specific dsp code, not e.g. all
uses of CONFIG_ and HAVE_ checks. But maybe it is already
enough to compile FFmpeg with MSVC with whole-programm-optimizations
enabled (if one does not disable too many components).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Y, U, V data is loaded at the end of the current iteration for the next
iteration.
It results in memory access past the frame data on the last iteration
(that data is never used after the loading).
So load data at the start of the iteration, so that only useful data is
loaded.
Signed-off-by: Vardan Margaryan <v.t.margaryan@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
This patch adds code to support specializations of the hscale function
and adds a specialization for filterSize == 4.
ff_hscale8to15_4_neon is a complete rewrite. Since the main bottleneck
here is loading the data from src, this data is loaded a whole block
ahead and stored back to the stack to be loaded again with ld4. This
arranges the data for most efficient use of the vector instructions and
removes the need for completion adds at the end. The number of
iterations of the C per iteration of the assembly is increased from 4 to
8, but because of the prefetching, there must be a special section
without prefetching when dstW < 16.
This improves speed on Graviton 2 (Neoverse N1) dramatically in the case
where previously fs=8 would have been required.
before: hscale_8_to_15__fs_8_dstW_512_neon: 1962.8
after : hscale_8_to_15__fs_4_dstW_512_neon: 1220.9
Signed-off-by: Jonathan Swinney <jswinney@amazon.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
This avoids having to rebuild big files every time FFMPEG_VERSION
changes (which it does with every commit).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This avoids unnecessary churn and build breakage for users, by
making sure the whole version.h is included like it has been so far,
while keeping the benefit of not needing to rebuild most files in
the ffmpeg tree on minor/micro bumps.
Signed-off-by: Martin Storsjö <martin@martin.st>
Also bump the minor versions of all libraries, to signify the
API change of splitting the version.h headers and adding the
new version_major.h header.
Signed-off-by: Martin Storsjö <martin@martin.st>
The range parameters need to be set up before calling
sws_init_context (which selects which fastpaths can be used;
this gets called by sws_getContext); solely passing them via
sws_setColorspaceDetails isn't enough.
This fixes producing full range YUV range output when doing
YUV->YUV conversions between different YUV color spaces.
Signed-off-by: Martin Storsjö <martin@martin.st>
Some of these were made possible by moving several common macros to
libavutil/macros.h.
While just at it, also improve the other headers a bit.
Reviewed-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Inside a function an unnecessary ';' is just a null statement;
yet outside of it it is actually illegal (but compilers happen
to accept it without warning except when using -pedantic).
So modify the macros to always expect the user to add a ';'.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is done a second time for 5.0 because master was
merged into 5.0 so that it contains the recent DOVI additions.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
In case of shared builds, some object files containing tables
are currently duplicated into other libraries: log2_tab.c,
golomb.c, reverse.c. The check for whether this is duplicated
is simply whether CONFIG_SHARED is true. Yet this is crude:
E.g. libavdevice includes reverse.c for shared builds, but only
needs it for the decklink input device, which given that decklink
is not enabled by default will be unused in most libavdevice.so.
This commit changes this by making it more explicit about what
to duplicate from other libraries. To do this, two new Makefile
variables were added: SHLIBOBJS and STLIBOBJS. SHLIBOBJS contains
the objects that are duplicated from other libraries in case of
shared builds; STLIBOBJS contains stuff that a library has to
provide for other libraries in case of static builds. These new
variables provide a way to enable/disable with a finer granularity
than just whether shared builds are enabled or not. E.g. lavd's
Makefile now contains: SHLIBOBJS-$(CONFIG_DECKLINK_INDEV) += reverse.o
Another example is provided by the golomb tables. These are provided
by lavc for static builds, even if one uses a build configuration
that makes only lavf use them. Therefore lavc's Makefile contains
STLIBOBJS-$(CONFIG_MXF_MUXER) += golomb.o, whereas lavf's Makefile
has a corresponding SHLIBOBJS-$(CONFIG_MXF_MUXER) += golomb_tab.o.
E.g. in case the MXF muxer is the only component needing these tables
only libavformat.so will contain them for shared builds; currently
libavcodec.so does so, too.
(There is currently a CONFIG_EXTRA group for golomb. But actually
one would need two groups (golomb_avcodec and golomb_avformat) in
order to know when and where to include these tables. Therefore
this commit uses a Makefile-based approach for this and stops
using these groups for the users in libavformat.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
