x86/flacdsp: add SSE2 and AVX decorrelate functions
Two to four times faster depending on instruction set, block size and channel count.
Showing
Please
register
or
sign in
to comment
Two to four times faster depending on instruction set, block size and channel count.