Could be that the driver actively polls a GPU device for status changes. Who knows. Maybe it's linked to the command queue. Then it could be an option to only have a OpenCL command queue generated when we are processing a pixelpipe and release it afterwards. Depends on the creation overhead if this makes sense. I'll give it a try.
Ulrich Am 17.02.2013 23:05, schrieb jeremy rosen: > maybe there is a (maybe amd specific) call to do to tell the GPU to go > to sleep for good ? > > On Sun, Feb 17, 2013 at 10:26 PM, johannes hanika <[email protected]> wrote: >> if you ask me, those guys just start 4 threads just in case. >> `MicroSleep' sounds like they're not up to anything useful at this >> time. maybe the call to ioctl indicates that they need to tell the >> device that nothing has happened? or maybe the device can request a >> cpu thread to wake up? >> ------------------------------------------------------------------------------ The Go Parallel Website, sponsored by Intel - in partnership with Geeknet, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials, tech docs, whitepapers, evaluation guides, and opinion stories. Check out the most recent posts - join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ darktable-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/darktable-devel
