libav.git - [no description]

	Commit message (Collapse)	Author	Age
*	SBR DSP: unroll sum_square	Christophe GISQUET	2012-03-07
\| \| \| \| \| \| \| \|	The length is even, so some unrolling can be performed. Timings are for x86: - 32bits: 102c -> 82c - 64bits: 82c -> 69c Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
*	SBR DSP x86: implement SSE sbr_sum_square_sse	Christophe GISQUET	2012-02-23
\| \| \| \| \| \| \| \| \| \| \| \| \|	The 32bits targets have been compiled with -mfpmath=sse for proper reference. sbr_sum_square C /32bits: 82c (unrolled)/102c C /64bits: 69c (unrolled)/82c SSE/32bits: 42c SSE/64bits: 31c Use of SSE4.1 dpps to perform the final sum is slower. Not unrolling to perform 8 operations in a loop yields 10 more cycles. Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
*	SBR DSP: use intptr_t for the ixh parameter.	Christophe GISQUET	2012-02-23
\| \| \| \|	Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
*	aacsbr: ARM NEON optimised sbrdsp functions	Mans Rullgard	2012-01-28
\| \| \| \| \| \|	Overall speedup of HE-AAC decoding 2.3x on Cortex-A8, 1.2x on A9. Signed-off-by: Mans Rullgard <mans@mansr.com>
*	aacsbr: move some simdable loops to function pointers	Mans Rullgard	2012-01-28
	This prepares for assembly optimisations by moving the most time-consuming loops to functions called through pointers in a new context. Signed-off-by: Mans Rullgard <mans@mansr.com>