-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 08/28/2011 06:51 PM, Bryan Cain wrote:
> On 08/28/2011 07:38 PM, Eric Anholt wrote:
>> On Sat, 27 Aug 2011 20:18:55 -0700, Kenneth Graunke
>> wrote:
>>> From: Bryan Cain
>>>
>>> Using multiply and reciprocal for integer division involves potentia
On Sun, 28 Aug 2011 20:51:10 -0500, Bryan Cain wrote:
> On 08/28/2011 07:38 PM, Eric Anholt wrote:
> > On Sat, 27 Aug 2011 20:18:55 -0700, Kenneth Graunke
> > wrote:
> >> From: Bryan Cain
> >>
> >> Using multiply and reciprocal for integer division involves potentially
> >> lossy floating point
On 08/28/2011 07:38 PM, Eric Anholt wrote:
> On Sat, 27 Aug 2011 20:18:55 -0700, Kenneth Graunke
> wrote:
>> From: Bryan Cain
>>
>> Using multiply and reciprocal for integer division involves potentially
>> lossy floating point conversions. This is okay for older GPUs that
>> represent integers
On Sat, 27 Aug 2011 20:18:55 -0700, Kenneth Graunke
wrote:
> From: Bryan Cain
>
> Using multiply and reciprocal for integer division involves potentially
> lossy floating point conversions. This is okay for older GPUs that
> represent integers as floating point, but undesirable for GPUs with
>
On 08/27/2011 10:18 PM, Kenneth Graunke wrote:
> From: Bryan Cain
>
> Using multiply and reciprocal for integer division involves potentially
> lossy floating point conversions. This is okay for older GPUs that
> represent integers as floating point, but undesirable for GPUs with
> native integer
From: Bryan Cain
Using multiply and reciprocal for integer division involves potentially
lossy floating point conversions. This is okay for older GPUs that
represent integers as floating point, but undesirable for GPUs with
native integer division instructions.
TGSI, for example, has UDIV/IDIV
TGSI, at the very least, has UDIV/IDIV instructions for integer division.
---
src/glsl/ir_optimization.h | 13 +++--
src/glsl/lower_instructions.cpp|4 +++-
src/mesa/drivers/dri/i965/brw_shader.cpp |1 +
src/mesa/program/ir_to_mesa.cpp|
