Should fix compilation with old Yasm/Nasm versions.
Signed-off-by: James Almer <jamrial@gmail.com>

~20% faster than AVX.
Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

This lets the cglobal macro automatically append a suffix to the function name.
This means that INIT_XMM avx must be used rather than INIT_AVX.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This reduces inter-library dependencies.

These are the only WMA bits shared with binkaudio. Splitting them off
reduces the binnkaudio dependency on general WMA code.

Code is only used during initialization, so malloc/free
should be fine to use.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Not sure it was meant to be used for something at some point though.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

A bit more complex than e.g. adding it to the context, but
using the context for something that will be used only during
initialization seemed a bit wasteful.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

avcodec/idctdsp: Initialize ff_put/add_pixels_clamped correctly so that the optimized functions are also used
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reset the settings as it is done for x264.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Thanks to Pascal Massimino and Michael Militzer for relicensing as LGPL.
Signed-off-by: Diego Biurrun <diego@biurrun.de>

These function pointers already existed in the ARM code. Adding them globally
allows calls to the function pointers to access arch-optimized versions of the
functions transparently.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

The HAVE_ prefix is reserved for macros set by configure.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Stops compiler from doing incredibly stupid things.
With vsynth1-flv inner loop goes from 3501 to 3275
decicycles.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

3646 -> 3597 decicycles in inner loop when decoding
vsynth1-flv.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Only set a value if _WIN32_WINNT is undefined or smaller than 0x0600. This is
cleaner than unconditional definition and avoids a number of redefinition
warnings. Also only define a value in one of the two dxva2 headers.

This avoids related incompatible pointer type warnings.

Together with the switch to RL_VLC this results in
a speedup of about 30% in this inner loop.
Overall speedup only relevant for medium to high bitrate
streams.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Together with the switch to RL_VLC this results in about
10% speedup for this inner loop.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Some additional optimizations in following patch.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Doesn't change generated code, just is more consistent
and a bit less error-prone.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>