These smaller samples do not need to be unpacked to double words
allowing the code to process more pixels every iteration (still 2 in MMX
but 6 in SSE2). It also avoids emulating the missing double word
instructions on older instruction sets.
Like with the previous code for 16-bit samples this has been tested on
an Athlon64 and a Core2Quad.
Athlon64:
1809275 decicycles in C, 32718 runs, 50 skips
911675 decicycles in mmx, 32727 runs, 41 skips, 2.0x faster
495284 decicycles in sse2, 32747 runs, 21 skips, 3.7x faster
Core2Quad:
921363 decicycles in C, 32756 runs, 12 skips
486537 decicycles in mmx, 32764 runs, 4 skips, 1.9x faster
293296 decicycles in sse2, 32759 runs, 9 skips, 3.1x faster
284910 decicycles in ssse3, 32759 runs, 9 skips, 3.2x faster
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This is a fairly dumb copy of the assembly for 8-bit samples but it
works and produces identical output to the C version. The options have
been tested on an Athlon64 and a Core2Quad.
Athlon64:
1810385 decicycles in C, 32726 runs, 42 skips
1080744 decicycles in mmx, 32744 runs, 24 skips, 1.7x faster
818315 decicycles in sse2, 32735 runs, 33 skips, 2.2x faster
Core2Quad:
924025 decicycles in C, 32750 runs, 18 skips
623995 decicycles in mmx, 32767 runs, 1 skips, 1.5x faster
406223 decicycles in sse2, 32764 runs, 4 skips, 2.3x faster
387842 decicycles in ssse3, 32767 runs, 1 skips, 2.4x faster
307726 decicycles in sse4, 32763 runs, 5 skips, 3.0x faster
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Since lavfi works natively with AVFrame now, these functions are no longer
necessary and can be removed in a future bump.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master:
af_join: do not leak input frames.
asrc_anullsrc: return EOF, not -1
Conflicts:
libavfilter/asrc_anullsrc.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'f4281f457194a6a4489fbd7423e2ab2e13c6d4d9':
asrc_anullsrc: do not set samplerate and channel layout explicitly
Conflicts:
libavfilter/asrc_anullsrc.c
No change to FFmpeg due to semantic difference between the filter sources
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'e4a7b2177d14678ae240edcabaacfe2b14619b7b':
vf_showinfo: remove its useless init function
AVOptions: fix using named constants with child contexts.
Conflicts:
libavutil/opt.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '9676b9a2cdc4a90611188fc48d8d388e427997c5':
AVOption: remove an unused function parameter.
filters.texi: restore mistakenly removed section name for noformat
avfiltergraph: use sizeof(var) instead of sizeof(type)
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '556aab8f11b045a21182eee32413aa78d5c8539b':
lavfi: use designated initializers in avfilter_class
Conflicts:
libavfilter/avfilter.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This is admittedly kind of pointless since usually -f image2pipe
can be used for the purpose, but this is more user-friendly.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Otherwise ffmpeg -formats claims that we only support demuxing
of that format.
To keep compatibility the struct could be duplicated instead,
but this seems almost like overkill for such a rare format.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
- Make MultiByteToWideChar fail when it encounters invalid encoding.
Without this, invalid characters might just be skipped
- When MultiByteToWideChar fails, assume the file name is in CP_ACP
and open it via normal open function, even when the file will be
written
- When malloc fails return error instead of crashing
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
To avoid allocating ridiculous amounts of memory for corrupted files,
read the input in chunks limited to filesize or an arbitrary large
amount when that is not known (chosen to be 50M).
* commit '9652d4fcfc9c07a726b35efc4ac644d9751b36d7':
ape: provide two additional bytes in buffer for old MAC versions
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'b164d66e35d349de414e2f0d7365a147aba8a620':
ape: make version-dependent decoding functions called via pointers
mpegts: add support for stream_type 0x42, which is CAVS
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '3e175a2a550f5c2f788e8cd969052e10a5025a27':
avconv: mark attachment streams as immediately finished
vaapi: fix argument for ff_vaapi_common_end_frame call
Merged-by: Michael Niedermayer <michaelni@gmx.at>
The check `src > dst' in the form `&c->out[-back] > c->out' invokes
pointer overflow, which is undefined behavior in C.
Remove the check. Also replace `&c->out[-back] < c->out_start' with
a safe form `c->out - c->out_start < back' to avoid overflow.
CC: libav-stable@libav.org
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Simplify `p->buf > p->buf + p->buf_size - 4' as `p->buf_size < 4'.
Avoid a possible out-of-bounds pointer, which is undefined behavior
in C.
CC: libav-stable@libav.org
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
When `off' is 0, `0x537F6103 << 32' in the following expression invokes
undefined behavior, the result of which is not necessarily 0.
(0x537F6103 >> (off * 8)) | (0x537F6103 << (32 - (off * 8)))
Avoid oversized shifting.
CC: libav-stable@libav.org
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
