Che: > I ran a 172.541 second test, my top 3 were: > merge_linear_argb_ref (/usr/lib/liboil-0.3.so.0.3.0) with 56.039 seconds > resample_linear_argb_ref (/usr/lib/liboil-0.3.so.0.3.0) with 39.708 seconds > yuv420p_to_rgb32 (/usr/lib/gstreamer-0.10/libgstffmpegcolorspace.so) > with 19.324 seconds > > This seems to be in line with your suggestion that liboil and > libgstffmpegcolorspace were the offenders ...I will add this detail to > the bugs you have filed.
Thanks. It's good to report these issues upstream, so hopefully the maintainers of the modules can address the problems. Also, since this issue seems to be affecting you, it might not be so hard to add some hardware acceleration yourself, then we could provide a patch to get upstream. If you have programming skills and an interest in learning about this, let me know offline and we can get you in touch with the right people at Sun who could help. > Many thanks for your guidance, I can categorically say that this > exercise has taught me something :) Great. I hope it taught you how easy it is to generate professional performance analysis reports on Solaris. On Solaris, this just works out of the box pretty easily as you notice, while on Linux using tools like valgrind an end-user probably would never be able to figure out how to generate a performance report like this. Brian
