On Wed, 2015-06-10 at 11:59 +0100, Chris Wilson wrote: > > sudo perf record -g -a -c 1 -e i915:i915_gem_request_wait_begin is > likely to capture the culprit. Though to be precise you need to look at > wait_begin + wait_end to measure the stall (some waits are more harmless > than others). As a first approximation just using perf will be good > enough to get a few leads.
Thanks Chris. But I no sooner posted that and a reboot later the problem went away and has not come back since. Good or bad depending on your perspective I suppose. Thanks for the response and suggestions though. Much appreciated. Cheers, b.
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
