On 08/29/2011 03:41 PM, Zack Rusin wrote:
> On Monday, August 29, 2011 04:02:08 PM Zack Rusin wrote:
>> Either way though if GL needs those ops then, like Brian mentioned, it'd be
>> a good idea to add them.
> Actually I think it seems easier to just flip the ops rather than add new
> instruction
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 08/29/2011 01:41 PM, Zack Rusin wrote:
> On Monday, August 29, 2011 04:02:08 PM Zack Rusin wrote:
>> Either way though if GL needs those ops then, like Brian mentioned, it'd be
>> a good idea to add them.
>
> Actually I think it seems easier to ju
On Monday, August 29, 2011 04:02:08 PM Zack Rusin wrote:
> Either way though if GL needs those ops then, like Brian mentioned, it'd be
> a good idea to add them.
Actually I think it seems easier to just flip the ops rather than add new
instructions, i.e. change stuff like a > b, to b < a, and a
On Monday, August 29, 2011 03:49:03 PM Bryan Cain wrote:
> I somehow didn't notice this until now, but why are there no Gallium
> "set" opcodes for integers that are equivalent to the SGT or SLE opcodes
> for floats? Does it have to do with available hardware features?
It's largely because D3D10
On Mon, Aug 29, 2011 at 1:49 PM, Bryan Cain wrote:
> I somehow didn't notice this until now, but why are there no Gallium
> "set" opcodes for integers that are equivalent to the SGT or SLE opcodes
> for floats? Does it have to do with available hardware features?
It's probably just an oversight
I somehow didn't notice this until now, but why are there no Gallium
"set" opcodes for integers that are equivalent to the SGT or SLE opcodes
for floats? Does it have to do with available hardware features?
Bryan
___
mesa-dev mailing list
mesa-dev@lists