Sindre Aamås 1995e03d91 [Processing/x86] Add an SSSE3 implementation of GeneralBilinearFastDownsample ... Keep track of relative pixel offsets and utilize pshufb to efficiently extract relevant pixels for horizontal scaling ratios <= 4. Fall back to a generic approach for ratios > 4. Note that the generic approach can be backported to SSE2. The implementation assumes that data beyond the end of each line, before the next line begins, can be dirtied; which AFAICT is safe with the current usage of these routines. Speedup is ~6.67x/~3.26x (32-bit/64-bit) for horizontal ratios <= 2, ~6.24x/~3.00x for ratios within (2, 4], and ~4.89x/~2.17x for ratios > 4 when not memory-bound on Haswell as compared with the current SSE2 implementation.		2016-05-23 20:23:31 +02:00
..
build/win32	Rename a vcproj folder to camelcase, to match all other folders in the same project	2015-03-25 11:46:41 +02:00
interface	Remove tabs in struct and class definitions	2015-06-10 10:22:01 +03:00
src	[Processing/x86] Add an SSSE3 implementation of GeneralBilinearFastDownsample	2016-05-23 20:23:31 +02:00
targets.mk	refine common moudle for part of intra prediction function	2014-09-25 14:03:11 +08:00