Re: yuvdeinterlace
stefan <[email protected]>
| Newsgroups | gmane.comp.video.mjpeg.devel |
|---|---|
| Message-ID | <[email protected]> |
Scott Smith schrieb: > On 2/23/07, stefan <[email protected]> wrote: > >> Could someone explain me, why it must be aligned to 16 bytes? 8 bytes I >> would understand (quad-word) but 16?!? Is this really necessary? >> > > I can't for Altivec, but I can for sse... > hmm, well, my sentence was not well formulated... My problem in understanding is not correlated to the potential speedup of certain opcodes used. It is more related to the fact, that all modern block-based-codecs use 8x8 pixel-blocks and if you do an overlapped search... OK, I can come arround that thing by using psad_sub22() instead, as this one just can not be forced on 16 byte boundaries... :-) (it is an 8x8 pixel-block-SAD...) But this will mean that there is no overlap anymore (I just do not know for sure if it will work reliably without...) cu Stefan ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV