On Wed, 2014-05-28 at 12:34 +0200, Torbjorn tyridal wrote: > I have the following case on a haswell 4770R (gigabyte brix) > > A libva based encoder reads raw 1920x1088 nv12 frames from file on > ramfs. Encoding set to 10 slices, fixed qp (26). cpu governor set to > performance.
> if I run two of the above in parallel, each goes along at 230fps, 18% > cpu (top). ie a total of 460 fps/36% cpu > If I then add a third instance the frame rate drops to 72fps (total 216) > and the cpu load jumps to 86 / 198%. > > Is there an explanation for the above in libva? I tested a 1920x1080 YUV file with cases avcenc on my HSW: 1 * 206 fps, 2 * 149 fps, 3 * 99 fs, 4 * 74 fps. What is the result if you set to 1 slice per frame ? Thanks Haihao > Decoder scales nicely: 1 * 926 fps, 2 * 626 fps, 3 * 411 fps, 4 * 288 > fps, 5 * 229 fps ... > > Cheers, T. > > _______________________________________________ > Libva mailing list > Libva@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/libva _______________________________________________ Libva mailing list Libva@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libva
