On 17/02/2018 11:36, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-02-15 15:34:53)
From: Tvrtko Ursulin <tvrtko.ursu...@intel.com>
A subtest to verify that the engine busyness is reported with expected
accuracy on platforms where the feature is available.
We test three patterns: 2%, 50% and 98% load per engine.
* Use spin batch instead of nop calibration.
* Various tweaks.
* Change loops to be time based.
* Use __igt_spin_batch_new inside timing sensitive loops.
* Fixed PWM sleep handling.
* Use restarting spin batch.
* Calibrate more carefully by looking at the real PWM loop.
* Made standalone.
* Better info messages.
* Tweak sleep compensation.
* Some final tweaks. (Chris Wilson)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursu...@intel.com>
Reviewed-by: Chris Wilson <ch...@chris-wilson.co.uk>
+ /* Sampling platforms cannot reach the high accuracy criteria. */
But we don't handle guc, right?
I'll dig up and rebase my old patch which implements busy stats in GuC
Or at least it doesn't work to sufficient accuracy. And bsw hung.
There are some occasional excursions over 15% tolerance even with
execlists on small core. Bummer. Don't want to be playing up the
tolerance game. I'll analyse in more detail and think what to do.
Do you have a link to BSW hang? Is that obviously related to PMU?
Intel-gfx mailing list