[vlc-devel] [PATCH] add ARM/NEON conversions for audio_filter/channel_mixer/simple

[Måns Rullgård]

David Geldreich <david.geldreich at free.fr> writes:

> write a ARM/NEON inline assembly version of most of the conversion cases of 
> audio_filter/channel_mixer/simple
>
> inline assembly is in separate functions for clarity and will be inlined by the compiler
>
> For example, 5.x->2 conversion gets a 8x speedup on iPad1 and 3x on iPad2
>
> I could provide a test program that shows that these routines :
> - give the same result (modulo epsilon) as the original one
> - work for any alignement of src/dst
> - work for any buffer size

Do yourself a favour and learn about the VMLA instruction.
This code can be made twice as fast.

-- 
Måns Rullgård
mans at mansr.com