[flac-dev] New routine: FLAC__lpc_compute_autocorrelation_asm_ia32_sse_lag_16

[Erik de Castro Lopo]

lvqcl wrote:

> libFLAC have three SSE-accelerated functions
> FLAC__lpc_compute_autocorrelation_asm_ia32_sse_lag_N (N = 4, 8, 12).
> They require lpc_order less than N.
> The best compression preset (flac -8) uses lpc_order up to 12; it means
> that during encoding FLAC also uses unaccelerated C function.
> 
> I'm not very familiar with asm so I took
> FLAC__lpc_compute_autocorrelation_asm_ia32_sse_lag_12, changed it and
> wrote function ..._sse_lag_16. According to my tests 'flac -8' encoding
> became 5% faster. The output files are identical to the output of 64-bit FLAC.
> 
> But as I said I'm not familiar with assembler so please check it.

Unfortunately this patch doesn't apply. How was it generated? Would it be
possible to send a patch generated by Git because those usually do work.

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