Re: [PATCH 5/5] d3d9/tests: Move a Present out of the loops in fog_with_shader_test().

[Henri Verbeet]

On 20 February 2015 at 13:06, Matteo Bruni <mbruni(a)codeweavers.com> wrote:
> It reduces the duration of the visual test with all the other functions
> from visual.c compiled out from 9 to 1.5 seconds for me (on a Radeon
> HD 5650M, Linux, Mesa).
> ---
> For what it's worth, I think we can actually just remove most of the
> Present() calls in the tests. It's perhaps fun to keep for nostalgic
> reasons, but I think the only time anyone really wants to see what
> gets drawn is while debugging failures.

[PATCH 5/5] d3d9/tests: Move a Present out of the loops in fog_with_shader_test().

[Matteo Bruni]

2015-02-20 13:33 GMT+01:00 Henri Verbeet <hverbeet(a)gmail.com>:
> On 20 February 2015 at 13:06, Matteo Bruni <mbruni(a)codeweavers.com> wrote:
> > It reduces the duration of the visual test with all the other functions
> > from visual.c compiled out from 9 to 1.5 seconds for me (on a Radeon
> > HD 5650M, Linux, Mesa).
> > ---
> > For what it's worth, I think we can actually just remove most of the
> > Present() calls in the tests. It's perhaps fun to keep for nostalgic
> > reasons, but I think the only time anyone really wants to see what
> > gets drawn is while debugging failures.

Yeah, I don't know. I just "fixed" the worst offender in my case.
We might want to keep just one present per test for the fun factor or
remove them all, either is fine to me.

This was mostly to show that those Present do take a lot of time, at
least on some HW configurations.