[Piglit] [PATCH 1/1] gles2: new test to render triangle and validate via glReadPixels
tom.gall at linaro.org
Mon Nov 26 11:25:53 PST 2012
Thanks for the feedback.
On Mon, Nov 26, 2012 at 10:59 AM, Eric Anholt <eric at anholt.net> wrote:
> Tom Gall <tom.gall at linaro.org> writes:
>> new gles2 test which very simply renders a triangle and validates
>> it was rendered correctly via reading from glReadPixels
>> and checking the returned buffer for values of alpha, red,
>> green and blue.
> Is a triangle really important for this test? This really looks like it
> ought to be a shader_runner test if you're trying to test minimal "can I
> render a primitive and read back?" and not the specific pixels lit by
> the triangle. But if you have shader_runner support then there are a
> ton of tests that are equivalent to rasterizing pixels with a constant
> color and reading back.
Yes this one is a sanity test. gles2 as you know has pretty much
nothing so, gotta start somewhere.
I haven't tried out shader_runner but yes, in general I'd rather adapt
than write whole new.
> PlainExecTest implies !runConcurrent, but why would you want
> non-concurrent anyway?
In my case I'm running on various ARM dev boards. Some of them are
quite modest when it comes to ram/cpu so yeah in this sanity testcase
and the others I'm working on there isn't anything that would prevent
"Where's the kaboom!? There was supposed to be an earth-shattering
kaboom!" Marvin Martian
Graphics Working Group | Linaro.org │ Open source software for ARM SoCs
w) tom.gall att linaro.org
h) tom_gall att mac.com
More information about the Piglit