flac dev - Jan 2014 - PATCH: asm versions for two _wide() functions

As I wrote earlier, GCC generates slow ia32 code for
FLAC__lpc_compute_residual_from_qlp_coefficients_wide()
and FLAC__lpc_restore_signal_wide(). So 24-bit encoding/decoding is slower
for GCC compile than for MSVS or ICC compile.

I took FLAC__lpc_compute_residual_from_qlp_coefficients_asm_ia32
and FLAC__lpc_restore_signal_asm_ia32 asm functions and wrote their _wide
versions.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: wide_asm.patch
Type: application/octet-stream
Size: 15653 bytes
Desc: not available
Url :
http://lists.xiph.org/pipermail/flac-dev/attachments/20140103/45594e52/attachment-0001.obj

lvqcl wrote:
> As I wrote earlier, GCC generates slow ia32 code for
FLAC__lpc_compute_residual_from_qlp_coefficients_wide()
> and FLAC__lpc_restore_signal_wide(). So 24-bit encoding/decoding is slower
> for GCC compile than for MSVS or ICC compile.
> 
> I took FLAC__lpc_compute_residual_from_qlp_coefficients_asm_ia32
> and FLAC__lpc_restore_signal_asm_ia32 asm functions and wrote their _wide
> versions.
Patch applied. Thanks.

I'l do a little more testing on this and the other patches before pushing
to git.

Cheers,
Erik
-- 
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/

Erik de Castro Lopo wrote:
> I'l do a little more testing on this and the other patches before
pushing
> to git.
According to my tests, the speed increase after the patch that changes
"call .get_eip0 / pop eax" to "call .mov_eip_to_eax / mov eax,
[esp] / ret"
is negligible or absent.

OTOH, libFLAC is a very wide-spread library, and it's better to do things
in a recommended way.