AMD RadeonSI Gallium3D Begins Simple CL Demos

[phoronix]

Phoronix: AMD RadeonSI Gallium3D Begins Simple CL Demos

The open-source AMD RadeonSI Gallium3D driver is beginning to work when it comes to running simple OpenCL programs on the Radeon HD 7000 series graphics cards...

http://www.phoronix.com/vr.php?view=MTM3ODM

[vadimg]

The open-source AMD OpenCL support is in even worse shape than the lacking and not-always-fast OpenGL support, but in the past few months for the HD 5000/6000 series hardware with the R600g driver it's begun to work for simple and small OpenCL programs

"In even worse shape than the lacking and not-always-fast OpenGL support" sounds really bad, as if nothing can be worse than OpenGL support in r600g. Really, Michael, are you trying to say that OpenGL support in r600g is the worst of all open-source drivers?

[duby229]

Beats the hell out of me too. My experience with r600g has been awesome. No glitches, no slowness, no tearing, no crashing, both monitors work. Better than catalyst. It only lack proper power management.

EDIT: Can someone explain the context in which r600g was mentioned? My understanding is that radeonsi and r600g are both mesa drivers one for HD2000 thru HD6000 and the other for GCN class HD7000 and up. The article is titled about radeonsi, but it seems that r600g was mentioned only to make fun of it.

[vadimg]

Maybe Michael should just forget about all sponsorship deals with any companies for a moment and publish e.g. Unigine Heaven 3.0 benchmark results for the top NVidia, AMD, and Intel GPUs in a single benchmark article, so that users could see what gpus perform better on open-source drivers...

[r1348]

EDIT: Can someone explain the context in which r600g was mentioned? My understanding is that radeonsi and r600g are both mesa drivers one for HD2000 thru HD6000 and the other for GCN class HD7000 and up. The article is titled about radeonsi, but it seems that r600g was mentioned only to make fun of it.

Nah, it was to shoehorn in a cross-reference to another article to potentially increase AD revenue. Phoronix always does that.

[ChrisXY]

Fyi with run_tests.sh from git://people.freedesktop.org/~tstellar/opencl-example

6 passes, 65 fails

And these seem to be rather trivial tests like

Running ./math-int add 1 2 3
Failed

LLVM 3.3 svn, mesa git master.

[tstellar]

Fyi with run_tests.sh from git://people.freedesktop.org/~tstellar/opencl-example

6 passes, 65 fails

And these seem to be rather trivial tests like

Running ./math-int add 1 2 3
Failed

LLVM 3.3 svn, mesa git master.

These all work for me on Evergreen, what GPU are you using?

[Michael]

"In even worse shape than the lacking and not-always-fast OpenGL support" sounds really bad, as if nothing can be worse than OpenGL support in r600g. Really, Michael, are you trying to say that OpenGL support in r600g is the worst of all open-source drivers?

No, not at all, it was in reference to performance against the proprietary drivers.

[duby229]

But nobody ever said that the OSS drivers would. It's been said repeatedly on these very forums that 70% of catalyst is about as close to the ideal as could be achieved. I'm certain that the open drivers have already passed that. The performance expectations were made perfectly clear from the beginning years ago. And it surpassed them. I don't understand where this nonsense about poor performance is coming from. Your very own benchmarks proved multiple times that the open drivers perform damn good.

I understand that many people play the open source shooters you like to use as benchmarks, but they don't stress modern hardware, therefore they -can't- represent what the hardware can do. It's just not possible with the games that you benchmark to represent performance on modern cards. You really -need- to update the game benchmarks to include modern day games that can stress modern day cards.

EDIT: Reach out to game developers and let them know that you need a solid stress test. PTS would be a kick ass way of showing off their stuff. It really is cool software. It just needs benchmarks that can stress modern GPU's. Talk to developers (or publishers) and let them know. Voice your need.

[smitty3268]

EDIT: Reach out to game developers and let them know that you need a solid stress test. PTS would be a kick ass way of showing off their stuff. It really is cool software. It just needs benchmarks that can stress modern GPU's. Talk to developers (or publishers) and let them know. Voice your need.

There are already Unigine tests in PTS, but Michael never seems to use them for some reason. And then there's a lot of possible games he could test through Steam now, but he doesn't seem interested in adding any of those, either.

What he's doing now isn't quite as bad as just running glxgears, but it's not much better.

I still don't think we've gotten a single article about how RadeonSI performs. It's enough to make me think Michael doesn't know how to compile all the different pieces to get it working. It's already up to basically GL3 support, now (through patches on the mailing list). Or maybe Michael just doesn't have any hardware?