[Piglit] [PATCH] cl: Add tests for mad mix
jan.vesely at rutgers.edu
Thu Oct 5 21:07:51 UTC 2017
On Thu, 2017-10-05 at 12:39 -0700, Matt Arsenault wrote:
> > On Oct 5, 2017, at 12:33, Jan Vesely <jan.vesely at rutgers.edu> wrote:
> > ah, that makes more sense. Do you mind if I add it to the commit
> > message?
you missed this part. I'd prefer if this purpose of the test was
included in the commit message. Otherwise the text a bit misleadingly
implies that it should also test half conversion.
> > (I'll also fix the formatting nits) with that
> > Reviewed-by: Jan Vesely <jan.vesely at rutgers.edu>
> > out of curiosity what's the use of having these in piglit? supposedly
> > the instruction selection and encoding part is tested in llvm lit. Is
> > this testing whether the instruction works correctly? shouldn't the hw
> > design team have tests for that?
> > Jan
> We can test an encoding in the lit tests, but we can’t actually check
> that it works. I don’t really trust the encoding tests until there’s
> something executing it. A lot of times in the past we’ve gotten the
> encodings wrong and the instruction doesn’t work, or the manual has
> had an off by one error in some of the encodings.
is the plan to add every gfx6+ instruction to piglit for that purpose?
if you have a generator for these tests wouldn't it make sense to
include it in piglit (and maybe drop gcn instruction tests from default
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 833 bytes
Desc: This is a digitally signed message part
More information about the Piglit