Added optimized AVX function for sample processing

Added AVX versions of existing 5 SSE functions. Added 6th AVX function to copy vectors which is 1.5 times faster then memcpy.
Data consistency and validness  is fully tested after processing with new AVX functions on aligned and non aligned buffers.
This commit is contained in:
Paul Davis 2015-05-12 21:07:09 -04:00
parent e2a76746e6
commit 6410aa896f
10 changed files with 75 additions and 25 deletions

View file

@ -145,6 +145,7 @@ def build(bld):
if bld.env['build_target'] == 'x86_64':
obj.defines += [ 'USE_X86_64_ASM' ]
if bld.env['build_target'] == 'mingw':
obj.defines += [ 'NO_POSIX_MEMALIGN' ]
obj.source += [ 'windows_special_dirs.cc' ]
obj.uselib += ' OLE'