The AMD Radeon R9 Fury Is Currently A Disaster On Linux

[eydee]

These articles are very good. While some butthurt AMD fans don't like them, this method seems to be the only thing inspiring AMD. Only when they get publicly ashamed enough are they willing to do anything about any bugs. Just think about the recent Bioshock infinite fiasco which took them half a year, a lot of people whining (including the developer) and many public criticism for something to happen. Keep up the good work and we might live in the days when AMD is usable under linux.

[oleid]

Considering the OpenCL performance, I wouldn't call it a disaster. OpenGL could be faster for sure. But I'd rather have legacy support by an open source driver than by an closed source blob with known bugs that hurt 2D desktop and don't get fixed (I'm looking at you, nvidia 340 driver series!)

[smitty3268]

It is not useless, particulary on higher resolutions... it is one (maybe only one in phoronix suite?) that clearly shows memory bandwidth difference And because it does that "up to bottom" for every chip tested

Why i want it? Well card has HBM isn't it

If you are interested in the pure hardware capabilities, just read one of the dozens of windows reviews.

Testing OpenArena on linux IS useless.

[duby229]

Now that we have several GCN 1.2 cards on the market, I think it's time you bite the bullet and get the AMDGPU driver up and running. Right now the only viable choice for video cards on linux is an nvidia card with closed source drivers, but I'm really curious to see if the amdgpu work could be a viable alternative.

Amdgpu is the kernel driver, hardware that uses that will stll use radeonsi for mesa. So I don't expect to see much difference in scalability.

[dungeon]

If you are interested in the pure hardware capabilities, just read one of the dozens of windows reviews.

Testing OpenArena on linux IS useless.

Nope it IS NOT useless, card has HBM that means better performance on nearly anything compared to current GDDR5 or slower... i like to see what that difference really is on most not biased possibile software, and i know openarena phoronix benchmark does that and i thought you will suggest me something else to run to test that, but no if you don't have other replacement for it is not useless

[Nobu]

70-75% of the performance of TITAN X (in games where the framerate never drops below 80fps) with a hell of a lot less memory and an unoptimized driver seems pretty decent to me. Only reason I wouldn't buy it is because I can't justify spending that much on just a GPU.

And, just to add to what I said earlier, that's 70-75% of the performance of a card that costs $400 (67%) more. Seems quite respectable, to me.

[duby229]

Nope it IS NOT useless, card has HBM that means better performance on nearly anything compared to current GDDR5 or slower... i like to see what that difference really is on most not biased possibile software, and i know openarena phoronix benchmark does that and i thought you will suggest me something else to run to test that, but no if you don't have other replacement for it is not useless

That's just not true. The bottleneck we are all seeing is obviously not at VRAM.

EDIT: It doesn't matter if you can get artificial conditions to show that bandwith if real actual usage conditions can't use it due to bottlenecks elsewhere.

[Serafean]

These articles are very good. While some butthurt AMD fans don't like them...

Call me a butthurt AMD fan, I don't care... I still think that some of the performance difference is because all the ports are mainly tested/developed on nvidia and their drivers. I will not dispute that Catalyst probably isn't nearly as good as it could be. The tests even Michael called neutral, are around 20%-30% difference(from top). Big Name games are around 50%. That difference is not within statistical error...

S.

[dungeon]

That's just not true. The bottleneck we are all seeing is obviously not at VRAM.

EDIT: It doesn't matter if you can get artificial conditions to show that bandwith if real actual usage conditions can't use it due to bottlenecks elsewhere.

You are wrong again, because openarena has fglrx profile so that should not be bottlenecked (unlike those steam games Michael tested whose has nvidia variabile, but not fglrx profile) ... maybe you are just not informed on that

[Michael]

Now that we have several GCN 1.2 cards on the market, I think it's time you bite the bullet and get the AMDGPU driver up and running. Right now the only viable choice for video cards on linux is an nvidia card with closed source drivers, but I'm really curious to see if the amdgpu work could be a viable alternative.

Evidently you didn't see my AMDGPU tests from several days ago - http://www.phoronix.com/vr.php?view=21911