[Intel-gfx] [PATCH] RFC drm/i915: Slaughter the thundering i915_wait_request herd

[Chris Wilson]

On Tue, Nov 03, 2015 at 10:03:19AM +0000, Chris Wilson wrote:
> On Tue, Nov 03, 2015 at 03:06:36AM +0000, Gong, Zhipeng wrote:
> > It seems that there are some gaps in the patch and first patch. 
> > Like there is no this line in the first patch. 
> >       if (req->ring->seqno_barrier)
> 
> Ah, that was in the context I hope...
> 
> > I have tried to apply this patch. And here is the cpu utilization and perf data on BDW
> 
> Do you also have a relative perf statistics like op/s we can compare to
> make sure we aren't just stalling the whole system?

Fwiw, I measure about a 3-5% (worst 20%) perf decrease in various X demos
with UXA when disabling the busywait for 20ms after we do an interrupt
driven wait. However, on the very same machine with the same benchmarks
SNA sees major improvement (which looks more likely to be from thermal
throttling affecting the results, but consistent enough to be very
interesting).
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre