Optimize floating-point celt_inner_prod() and dual_inner_prod() for ARM NEON
authorLinfeng Zhang <linfengz@google.com>
Wed, 25 Jan 2017 17:45:48 +0000 (09:45 -0800)
committerJean-Marc Valin <jmvalin@jmvalin.ca>
Tue, 6 Jun 2017 21:12:20 +0000 (17:12 -0400)
commit4fe9de3b0f17f4d6146a3d541daf869742629c5d
tree5f75241d09846e089cd8d015a3458e02cc4cd372
parent4d4ed208b1dfd8cdbb1df89382fbaa1815972e55
Optimize floating-point celt_inner_prod() and dual_inner_prod() for ARM NEON

The floating-point optimizations are not bit exact with C functions,
because of the different orders of floating-point operations.
But they are bit exact with the simulation C functions which simulate
the floating operations in the optimizations.

Change-Id: I149fda5b602fd5712b16fc8983df3c6c0c9e76ad

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>
celt/arm/arm_celt_map.c
celt/arm/pitch_arm.h
celt/arm/pitch_neon_intr.c
celt/pitch.h