[Intel-gfx] [PATCH] drm/i915: Restore waitboost credit to the synchronous waiter

Eero Tamminen eero.t.tamminen at intel.com
Tue Dec 8 00:35:56 PST 2015


On 12/07/2015 07:38 PM, Jesse Barnes wrote:
> Still wishing we had a good way to benchmark these types of changes
> across a range of workloads.  Eero, have you guys looked at turbo stuff
> at all yet?

Except for this, not really:

To make results comparable other platforms, I'm using sustained 
performance numbers by discarding results for first round(s) that were 
run with turbo.

For some of performance tracks, we've also planned to disable power 
management completely (e.g. by using fixed freqs for CPU & GPU) because 
it causes way too much variance to performance results (more between 
runs, than within single run).

	- Eero

> Also, is the boost logic only documented in misc commit messages?  Or do
> we have a nice block of text somewhere describing the intent (which may
> not match our implementation!) and how we try to achieve it?
> Those are both new requests though, so no need to block this patch:
> Reviewed-by: Jesse Barnes <jbarnes at virtuousgeek.org>

More information about the Intel-gfx mailing list