Modern Intel Gallium3D Driver Still Being Toyed With

[phoronix]

Phoronix: Modern Intel Gallium3D Driver Still Being Toyed With

While it's not the default Linux graphics driver for Sandy Bridge or Ivy Bridge hardware, the "ilo" independently-developed Gallium3D driver for modern Intel graphics hardware continues to be developed...

http://www.phoronix.com/vr.php?view=MTM3NDg

[curaga]

How nice to have a driver named "happiness" or "joy". Though that does put up some high expectations :P

[duby229]

Where is the comments about missing features and poor performance and the insistence that only the classic driver be used?

Oh never mind I get it. Bias only works one way.

And how come We have never seen an Intel APU vs AMD APU comparison? Somehow I doubt that will ever happen on this site. Oh sure there have been some that included benchmarks of both, but not t a single one of them included graphics benches. Which is arguably the most important part of an APU. Why the hell else would you buy an APU?

Either way. I don't care if you take a negative bias towards everything or if you take a positive bias towards everything. Just pick a damn side and be fair to that side.

http://www.phoronix.com/scan.php?pag...10_5800k&num=1
Review of the 5800k No Intel benches, but includes graphics.

http://www.phoronix.com/scan.php?pag...a8_3870k&num=3
Review of the 3870k Intel Benches, but no graphics.

If that isnt completely obvious. I mean come on. This is ridiculous.

[artivision]

Where is the comments about missing features and poor performance and the insistence that only the classic driver be used?

Oh never mind I get it. Bias only works one way.

And how come We have never seen an Intel APU vs AMD APU comparison? Somehow I doubt that will ever happen on this site. Oh sure there have been some that included benchmarks of both, but not t a single one of them included graphics benches. Which is arguably the most important part of an APU. Why the hell else would you buy an APU?

Either way. I don't care if you take a negative bias towards everything or if you take a positive bias towards everything. Just pick a damn side and be fair to that side.

http://www.phoronix.com/scan.php?pag...10_5800k&num=1
Review of the 5800k No Intel benches, but includes graphics.

http://www.phoronix.com/scan.php?pag...a8_3870k&num=3
Review of the 3870k Intel Benches, but no graphics.

If that isnt completely obvious. I mean come on. This is ridiculous.

There are benchmarks that show that Intel HD4000 is as fast as an Radeon 6500-600 Apu with 500Gflops and Catalyst. Any way we have not a reason to test a product (Catalyst) that steal to any Benchmark, reducing precision. I just don't understand how from this: http://www.xbitlabs.com/articles/gra...k_5.html#sect3 that Unigine shows the truth (gtx580=70% more Flops than HD6970, the same is with HD7000-3.8Tflops vs 6.4Tflops of a GTX680), now are equal.

[duby229]

Those are high end discrete cards and not APUs. Needless to say Intel doesn't make high end discrete cards.

[Andrecorreia]

???

what s point for this drivers? someone explain?

[duby229]

To unify all the oss drivers onto the gallium3d stack.

[mattst88]

what s point for this drivers? someone explain?

It's just a toy.

To unify all the oss drivers onto the gallium3d stack.

No, it's a toy, as acknowledged by its author.

[smitty3268]

i915g is supposedly faster now

Separately, for those curious about the recent prospects of the older i915 Gallium3D driver possibly becoming the default over Intel's i915 classic driver for supporting up through i945 IGPs, there isn't any new information.

Well, there was actually a few commits by that google developer which supposedly increased performance of the i915g driver by like 30+%.

[Awesomeness]

Intel ended up bringing OpenGL 2.1 to classic i915 to compete with that older Gallium3D driver

I don't get Intel?s total refusal to support Gallium3D.

I understand that if there is a fine working classic driver for older hardware, the GPU vendor would not want to spend money on a Gallium port (AMD did that as well). But considering that i915g exists already and reportedly is better than the classic driver, to then spend the resources on the classic driver to achieve partial feature parity with the Gallium driver makes no economic sense from my POV. Shouldn't Intel?s priority be to make their hardware customers as satisfied as possible to make them want to buy Intel products as well?

Anybody got a clue why Intel insists on not touching Gallium ? not even for new drivers?