Hi,<br><br>Thank you for your commit.<br><br><div><span class="gmail_quote">2007/11/19, Guillaume POIRIER <<a href="mailto:poirierg@gmail.com">poirierg@gmail.com</a>>:<br><br></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> o pixel_ssd_8x8_altivec()<br><br>As far as this once is concerned, I seem to be too sleepy to find a<br>set of command-line options that trigger this usage. Would you have<br>one that I could use for testing?<br><br>
Not committed.</blockquote><div><br> This function is called at<br><br>encoder/rdo.c:<br><br>int ssd_mb( x264_t *h )<br>{<br> return h->pixf.ssd[PIXEL_16x16]( h->mb.pic.p_fenc[0], FENC_STRIDE,<br> h->
mb.pic.p_fdec[0], FDEC_STRIDE )<br>* + h->pixf.ssd[PIXEL_8x8]( h->mb.pic.p_fenc[1], FENC_STRIDE,<br> h->mb.pic.p_fdec[1], FDEC_STRIDE )<br>* + h->pixf.ssd[PIXEL_8x8]( h->
mb.pic.p_fenc[2], FENC_STRIDE,<br> h->mb.pic.p_fdec[2], FDEC_STRIDE );<br>}<br><br>ssd_mb() is called at<br><br>o x264_rd_cost_mb() ( m>=6 )<br>o x264_macroblock_analyse() ( B-Frame, brdo, m>=6 )
<br><br>so I use the following options.<br></div><br>--bframes 3 --ref 5 -m 7 --b-rdo<br><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
BTW, If you're looking for some more stuff to optimize, maybe you<br>could have a look at the macro<br>ALTIVEC_STORE_SUM_CLIP: the 2 known cases where this macro is used is<br>with 8-byte aligned memory or 16-bytes aligned memory, so it should be
<br>possible to write a specialized version for these two cases, that<br>would be a bit faster.<br><br>This has been discussed already in the thread:<br><a href="http://www.via.ecp.fr/via/ml/x264-devel/2006-09/msg00068.html">
http://www.via.ecp.fr/via/ml/x264-devel/2006-09/msg00068.html</a><br></blockquote></div><br>OK. I'll see them.<br><br>