AMD Catalyst 14.6 Does Slightly Better With APITest OpenGL Tests

[lowflyer]

Palm Grease

... It seems for some reasons you like AMD bashing. ...

Some months ago Michael's articles used to be a lot more neutral. It changed when AMD stopped to send samples down his address. He has since become a complete AMD basher. To me this looks like Michael is "buyable". Just send a couple of samples to his address and he will start praising...
Just look how he does it up and down the line for all the Intel products which have not yet been released.

[_ONH_]

Because AMD PR requested new benchmarks be done from the new driver given the last "slaughtering" article given the new driver...

you did not unterstand what they told you. they meant old vs new driver comparaisson, like you tell in the Artikel heading. I was surprised that a Geforce Titan has something to do with beter Radeon driver. It seems for some reasons you like AMD bashing. Or what is te reason do do no driver version comparaisson in an article which the title sugest this, but instead the content just displays how Nv is faster in this nv benchmark.

[mibo]

I'm NOT saying that results are biased or incorrect. Just their usefulness is dubious as previous results could as way be reposted (for AMD...)

especially, since the title of the article is "AMD Catalyst 14.6 Does Slightly Better With APITest OpenGL Tests"...
Then it doesn't compare old vs. new driver and has 3 AMD cards and 7 NVIDIA cards in it

[przemoli]

Michael You really need an option to compare current test results with similar in the past.

As its now. You tested ONLY vendor vs vendor.

Not even incremental change between previous AMD driver and current one.
(And current driver is old one, anyway, as AMD had troubles with real 14.6 so it needed to be pulled back :|, so Yes, those changes Graham refer to? They aren't there)

Not even scaling between different solutions for same problem.


I'm NOT saying that results are biased or incorrect. Just their usefulness is dubious as previous results could as way be reposted (for AMD...)

[asdfblah]

Just tell them the open source driver is way better than fglrx , and that they should ditch fglrx completely and move to the open source driver in linux... :P

[schmidtbag]

Because AMD PR requested new benchmarks be done from the new driver given the last "slaughtering" article given the new driver...

I greatly appreciate your honesty and bluntness. While I wholeheartedly support AMD's recent efforts toward linux, if they don't want such harsh criticisms, they need to step things up. Look at nvidia - a couple years ago, they had some pretty faulty drivers in some cases, they had no intention on supporting anything open source, and they had no intention on supporting optimus. Moments like Linus' flipping-off-nvidia moment, Valve gaining interest in linux, and websites like this pointing out serious flaws in nvidia drivers have forced nvidia to completely turn things around. All we need now is some dedicated work toward nouveau.

But, with the rate AMD is going, we won't be needing catalyst too much longer. At this point, all I really want is some polished openCL support and crossfire support - everything else is nice but currently not holding me back.

[dungeon]

Does APITest also work with the open drivers ? It'd be interesting to see how they compare in such "exotic" tests against closed source stuff.

No, these tests needs GLSL 4.10 and some extensions for some of the tests are i guess only available in nvidia implementation .

[mazumoto]

Does APITest also work with the open drivers ? It'd be interesting to see how they compare in such "exotic" tests against closed source stuff.

[_SXX_]

Unfortunately he does not write on twitter anymore (... or he is on a loooong holiday).

I miss his posts about OpenGL...

Might be he just busy with SIGGRAPH.

[entropy]

I think it would have been interesting to see the difference between current and previous catalyst, to see the changes.

I guess it was exactly this what AMD PR actually intended by the request.