Timothy B. Terriberry authored 11 years ago and Jean-Marc Valin committed 11 years ago

Optimizing celt_pitch_xcorr()/xcorr_kernel() which also speeds up
FIRs, IIRs and auto-correlations

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

With gcc-4.4 at least, the raw asm.s files will always successfully
compile even if the default -march for the compiler would not support
those instructions.  So switch to testing the inline asm versions,
where the compiler will barf if they aren't supported by the default
arch if no -march is explicitly given, or if they aren't supported by
the requested -march when it is.

Original patch by Aurélien Zanelli <aurelien.zanelli@parrot.com>:
 http://lists.xiph.org/pipermail/opus/2013-May/002078.html

Revised version:
- Add autconf detection (ported from libtheora).
- Rename ARM5E to ARMv5E (an ARM5 is not the same thing as ARMv5!).
- Use actual macros so they can still be selectively overridden.
- Split out ARMv4 parts and add a few more ARMv4 macros.
- Label blocks to make them easy to find in generated assembly.
- Fix MULT16_32_Q15() so we can pass make check.
  The MDCT test passes in values larger than 2**30 for b.
  The new version should be just as fast (or faster, since it's
   easier to merge the shift with following instructions), and
   there's no appreciable impact on accuracy (FFT/MDCT SNR actually
   goes up in most cases).
- Fix register constraints.
  We were using early-clobber flags in a bunch of places that
   didn't need them, and commutative-pair flags in a bunch of
   places that weren't actually commutative.
  This was Jean-Marc's fault (the original code came from Speex).
- Simplify silk_CLZ16().
- Port over iFFT C_MULC asm by Andree Buschmann
   <AndreeBuschmann@t-online.de> from Rockbox.
- Speed up the C_MULC asm by using LDRD, allowing more flexible
   addressing, re-ordering instructions to avoid some stalls,
   allowing more flexible register allocation, and getting things
   out of the inline asm block so the compiler can schedule them
   better.
- Add C_MUL and C_MUL4 asm for the FFT to the encoder based, on the
   new C_MULC.

In total, this patch gives a 22.3% speed-up on test_opus_encoder on
 a 600 MHz Cortex A8 using gcc 4.2.1,
When restricted to ARMv4 optimizations, it gives a 9.6% speed-up
 on the same processor/compiler.
On the conformance test vectors:
 Average mono quality is 97.0583 %
 Average stereo quality is 97.775 %