On Fri, May 15, 2015 at 10:32:44AM -0700, Jason Ekstrand wrote:
> Hey Topi! Thanks for working on this!
>
> I've been meaning to do something for fp16 and fp64 in NIR for a while
> and just haven't gotten to it. I just sent out three "patches" laying
> out how I was intending to do it. I put "p
Hey Topi! Thanks for working on this!
I've been meaning to do something for fp16 and fp64 in NIR for a while
and just haven't gotten to it. I just sent out three "patches" laying
out how I was intending to do it. I put "patches" in quotes because
they're so half-baked they barely compile. It's
On Fri, May 15, 2015 at 2:39 AM, Topi Pohjolainen
wrote:
> I wanted to kick-off discussion on how to support floating point
> precision qualifiers in NIR. This is purely on optimization for
> GLES where one can reduce the number of GPU cycles. At the moment
> the compiler discards the qualifiers e
On Fri, May 15, 2015 at 12:32:52PM -0400, Rob Clark wrote:
> On Fri, May 15, 2015 at 12:22 PM, Pohjolainen, Topi
> wrote:
> > On Fri, May 15, 2015 at 11:59:25AM -0400, Rob Clark wrote:
> >> On Fri, May 15, 2015 at 5:39 AM, Topi Pohjolainen
> >> wrote:
> >> > I wanted to kick-off discussion on how
On Fri, May 15, 2015 at 12:22 PM, Pohjolainen, Topi
wrote:
> On Fri, May 15, 2015 at 11:59:25AM -0400, Rob Clark wrote:
>> On Fri, May 15, 2015 at 5:39 AM, Topi Pohjolainen
>> wrote:
>> > I wanted to kick-off discussion on how to support floating point
>> > precision qualifiers in NIR. This is pu
On Fri, May 15, 2015 at 11:59:25AM -0400, Rob Clark wrote:
> On Fri, May 15, 2015 at 5:39 AM, Topi Pohjolainen
> wrote:
> > I wanted to kick-off discussion on how to support floating point
> > precision qualifiers in NIR. This is purely on optimization for
> > GLES where one can reduce the number
On Fri, May 15, 2015 at 5:39 AM, Topi Pohjolainen
wrote:
> I wanted to kick-off discussion on how to support floating point
> precision qualifiers in NIR. This is purely on optimization for
> GLES where one can reduce the number of GPU cycles. At the moment
> the compiler discards the qualifiers e
I wanted to kick-off discussion on how to support floating point
precision qualifiers in NIR. This is purely on optimization for
GLES where one can reduce the number of GPU cycles. At the moment
the compiler discards the qualifiers early when abstract syntax
tree (AST) is transformed into intermedi
