On 11/05/2018 09:31, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-05-11 09:25:00)
On 10/05/2018 18:40, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-05-10 18:26:31)
On 10/05/2018 17:25, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-05-10 17:09:14)
On 09/05/2018 15:27, Chris
Quoting Tvrtko Ursulin (2018-05-11 09:25:00)
>
> On 10/05/2018 18:40, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2018-05-10 18:26:31)
> >>
> >> On 10/05/2018 17:25, Chris Wilson wrote:
> >>> Quoting Tvrtko Ursulin (2018-05-10 17:09:14)
>
> On 09/05/2018 15:27, Chris Wilson wrote:
>
On 10/05/2018 18:40, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-05-10 18:26:31)
On 10/05/2018 17:25, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-05-10 17:09:14)
On 09/05/2018 15:27, Chris Wilson wrote:
Bypass using the tasklet to submit the first request to HW, as the
tasklet
Quoting Tvrtko Ursulin (2018-05-10 18:26:31)
>
> On 10/05/2018 17:25, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2018-05-10 17:09:14)
> >>
> >> On 09/05/2018 15:27, Chris Wilson wrote:
> >>> Bypass using the tasklet to submit the first request to HW, as the
> >>> tasklet may be deferred unto
On 10/05/2018 17:25, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-05-10 17:09:14)
On 09/05/2018 15:27, Chris Wilson wrote:
Bypass using the tasklet to submit the first request to HW, as the
tasklet may be deferred unto ksoftirqd and at a minimum will add in
excess of 10us (and maybe tens
Quoting Tvrtko Ursulin (2018-05-10 17:09:14)
>
> On 09/05/2018 15:27, Chris Wilson wrote:
> > Bypass using the tasklet to submit the first request to HW, as the
> > tasklet may be deferred unto ksoftirqd and at a minimum will add in
> > excess of 10us (and maybe tens of milliseconds) to our
On 09/05/2018 15:27, Chris Wilson wrote:
Bypass using the tasklet to submit the first request to HW, as the
tasklet may be deferred unto ksoftirqd and at a minimum will add in
excess of 10us (and maybe tens of milliseconds) to our execution
latency. This latency reduction is most notable when
Bypass using the tasklet to submit the first request to HW, as the
tasklet may be deferred unto ksoftirqd and at a minimum will add in
excess of 10us (and maybe tens of milliseconds) to our execution
latency. This latency reduction is most notable when execution flows
between engines.
v2: Beware