1) each of the loops run within a single CTB, so the relevant reference
list is constant
2) when that CTB is, or lies on the same slice as, the current one, we
can use a simple access instead of a relatively expensive call to
ff_hevc_get_ref_list()
It allows attaching other external, opaque data to the frame and passing it
through the reordering process, for cases when the caller wants other data
than just the plain packet pts. There is no way to cleanly achieve this
without the field.
The input data must remain constant, make a copy instead. This is in
theory a performance hit, but since I failed to find any samples
using this feature, this should not matter in practice.
Also, check the size of the header, avoiding invalid reads on truncated
data.
CC:libav-stable@libav.org
This tries to find the most expressive part of the output of
armcc --vsn to include, giving a compiler identification of
"ARM Compiler 5.04 update 2 (build 82)" instead of just
"ARM Compiler 5.04" for armcc 5.0.
4.x versions of armcc output the following, for "armcc --vsn":
ARM C/C++ Compiler, RVCT4.0 [Build 925]
For evaluation purposes only
Software supplied by: ARM Limited
ARM C/C++ Compiler, 4.1 [Build 894]
For evaluation purposes only
Software supplied by: ARM Limited
5.0 versions output this:
Product: ARM Compiler 5.04
Component: ARM Compiler 5.04 update 2 (build 82)
Tool: armcc [5040081]
For evaluation purposes only
Software supplied by: ARM Limited
Signed-off-by: Martin Storsjö <martin@martin.st>
llvm's integrated assembler supports the AArch64 asm on darwin since
August 2014. So check $as first before using gas-preprocessor.pl via
$gas. Makes the checks specific for that the architecture specific asm
needs. PPC Altivec and AArch64 needs on ':vararg' for macro arguments.
Arm needs in addition the '.altmacro' directive.
The previous implementation of the parser made four passes over each input
buffer (reduced to two if the container format already guaranteed the input
buffer corresponded to frames, such as with MKV). But these buffers are
often 200K in size, certainly enough to flush the data out of L1 cache, and
for many CPUs, all the way out to main memory. The passes were:
1) locate frame boundaries (not needed for MKV etc)
2) copy the data into a contiguous block (not needed for MKV etc)
3) locate the start codes within each frame
4) unescape the data between start codes
After this, the unescaped data was parsed to extract certain header fields,
but because the unescape operation was so large, this was usually also
effectively operating on uncached memory. Most of the unescaped data was
simply thrown away and never processed further. Only step 2 - because it
used memcpy - was using prefetch, making things even worse.
This patch reorganises these steps so that, aside from the copying, the
operations are performed in parallel, maximising cache utilisation. No more
than the worst-case number of bytes needed for header parsing is unescaped.
Most of the data is, in practice, only read in order to search for a start
code, for which optimised implementations already existed in the H264 codec
(notably the ARM version uses prefetch, so we end up doing both remaining
passes at maximum speed). For MKV files, we know when we've found the last
start code of interest in a given frame, so we are able to avoid doing even
that one remaining pass for most of the buffer.
In some use-cases (such as the Raspberry Pi) video decode is handled by the
GPU, but the entire elementary stream is still fed through the parser to
pick out certain elements of the header which are necessary to manage the
decode process. As you might expect, in these cases, the performance of the
parser is significant.
To measure parser performance, I used the same VC-1 elementary stream in
either an MPEG-2 transport stream or a MKV file, and fed it through avconv
with -c:v copy -c:a copy -f null. These are the gperftools counts for
those streams, both filtered to only include vc1_parse() and its callees,
and unfiltered (to include the whole binary). Lower numbers are better:
Before After
File Filtered Mean StdDev Mean StdDev Confidence Change
M2TS No 861.7 8.2 650.5 8.1 100.0% +32.5%
MKV No 868.9 7.4 731.7 9.0 100.0% +18.8%
M2TS Yes 250.0 11.2 27.2 3.4 100.0% +817.9%
MKV Yes 149.0 12.8 1.7 0.8 100.0% +8526.3%
Yes, that last case shows vc1_parse() running 86 times faster! The M2TS
case does show a larger absolute improvement though, since it was worse
to begin with.
This patch has been tested with the FATE suite (albeit on x86 for speed).
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Initialise VC1DSPContext for parser as well as for decoder.
Note, the VC-1 code doesn't actually use the function pointer yet.
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
