On Thu, 2021-03-18 at 14:39 +0200, Ville Syrjälä wrote:
> On Thu, Mar 18, 2021 at 08:39:01AM +0200, Martin Peres wrote:
> > On 18/03/2021 00:45, Lyude wrote:
> > > From: Lyude Paul
> > >
> > > Since pre-nve4 only has two cursor sizes (32x32 and 64x64), we should at
> > > least test both of them.
On Thu, Mar 18, 2021 at 08:39:01AM +0200, Martin Peres wrote:
> On 18/03/2021 00:45, Lyude wrote:
> > From: Lyude Paul
> >
> > Since pre-nve4 only has two cursor sizes (32x32 and 64x64), we should at
> > least test both of them.
>
> This adds 36 subtests, which take about 1s in average. So the
On Thu, Mar 18, 2021 at 08:39:01AM +0200, Martin Peres wrote:
> On 18/03/2021 00:45, Lyude wrote:
> > From: Lyude Paul
> >
> > Since pre-nve4 only has two cursor sizes (32x32 and 64x64), we should at
> > least test both of them.
>
> This adds 36 subtests, which take about 1s in average. So the
On 18/03/2021 00:45, Lyude wrote:
From: Lyude Paul
Since pre-nve4 only has two cursor sizes (32x32 and 64x64), we should at
least test both of them.
This adds 36 subtests, which take about 1s in average. So the runtime is
not significantly increased on the Intel side.
It also seems that