[Mesa-dev] [PATCH] r600g: Implement GL_ARB_texture_gather

[Marek Olšák]

My advice is: The shader being compiled should be analyzed before
compilation and an array of bool flags (or one uint16_t for all flags)
should be produced saying which sampler units use TG4. Then, you only
need to set the swizzle in the key for the i-th sampler if the i-th
sampler is really used by TG4.

The shader analysis can be done in tgsi_scan_shader, which should be
pretty trivial.

However, this won't work with GL4 sampler arrays and indirect
addressing of those arrays, so.... well, if I were you, I would
probably ignore the ZERO and ONE swizzle cases and just put up with
the fact some piglit tests will always fail. That's the only use of
the tg4_swizzle array as far as I can see, so it can be removed.

Marek

On Mon, Jul 14, 2014 at 11:27 PM, Dave Airlie <airlied at gmail.com> wrote:
>>> Also, have you looked at Dave Airlie's impl? Not sure what's been
>>> going on there...
>>> http://cgit.freedesktop.org/~airlied/mesa/log/?h=r600g-texture-gather
>>> . He seemed to enable 4 components for >= CEDAR.
>>>
>>
>> No, didn't pop up when i searched the archives for any prior work on this.
>> Interesting! It looks like the GL_ARB_texture_gather portion in that branch
>> is almost exactly equivalent to my patch.
>>
>> It also has some of the additional parts needed for GL_ARB_gpu_shader5 such
>> as non-constant offsets, but i don't see where it would trigger shader
>> recompiles if texture sampler swizzle changes? Does it pass piglit?
>>
>>
>> David, any opinions on how to move forward with this feature?
>
> I didn't like my implementation sufficiently to push it forward, but I
> think the main thing
> was the extra key changes, I meant to get Marek's input on that code,
> but then it fell
> off my radar. So get input from Marek, repost and merge :)
>
> Dave.
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev