From: Ian Romanick
Just like the non-constant array index lowering pass, compare all N
indices at once. For accesses to a vec4, this saves 3 comparison
instructions on a vector architecture.
---
src/glsl/lower_vec_index_to_cond_assign.cpp | 57 --
1 files changed, 35 i
On Thu, 21 Jul 2011 12:16:58 -0700, "Ian Romanick" wrote:
> From: Ian Romanick
>
> Just like the non-constant array index lowering pass, compare all N
> indices at once. For accesses to a vec4, this saves 3 comparison
> instructions on a vector architecture.
> ---
> src/glsl/lower_vec_index_to
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 07/21/2011 05:28 PM, Eric Anholt wrote:
> On Thu, 21 Jul 2011 12:16:58 -0700, "Ian Romanick"
> wrote:
>> From: Ian Romanick
>>
>> Just like the non-constant array index lowering pass, compare all N
>> indices at once. For accesses to a vec4, thi
On Fri, 22 Jul 2011 09:27:36 -0700, Ian Romanick wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> On 07/21/2011 05:28 PM, Eric Anholt wrote:
> > On Thu, 21 Jul 2011 12:16:58 -0700, "Ian Romanick"
> > wrote:
> >> From: Ian Romanick
> >>
> >> Just like the non-constant array index lo
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 07/22/2011 11:27 AM, Eric Anholt wrote:
> On Fri, 22 Jul 2011 09:27:36 -0700, Ian Romanick wrote:
>> -BEGIN PGP SIGNED MESSAGE-
>> Hash: SHA1
>>
>> On 07/21/2011 05:28 PM, Eric Anholt wrote:
>>> On Thu, 21 Jul 2011 12:16:58 -0700, "Ian Roma
On Fri, 22 Jul 2011 15:20:03 -0700, Ian Romanick wrote:
> I'd like to refactor that code to be more like the non-constant array
> index lowering pass. The r-value and l-value case only differ in the
> order of the operands in the generated assignments and the optional
> if-statement wrapping. It