12-Way RadeonSI OpenGL Comparison vs. NVIDIA On Ubuntu Linux

[ua=42]

Is the point of these at 1080p to show which cards/drivers are bottlenecking the most at 1080p? There is some interesting situations where the GTX 980 beats the 1080 cards, not sure whats going on there?!

1080p is probably the most common screen size. So its useful on knowing... how well will my game run. Which for the older games is "fast enough" But we still get useful info for some of the more demanding benchmarks and games.
4K just plain doesn't work on many cards because they don't have enough ram to hold the textures.

[torsionbar28]

4K just plain doesn't work on many cards because they don't have enough ram to hold the textures.

Plus 4k monitors are still pretty spendy, and even more spendy when you realize a top-tier GPU is needed to drive it, making 4k somewhat rare still in mainstream gaming.

[fa5hion]

Phoronix: 12-Way RadeonSI OpenGL Comparison vs. NVIDIA On Ubuntu Linux

After posting a number of NVIDIA GPU Linux benchmarks this week using their latest drivers, here is similar treatment on the Radeon side using their newest open-source driver code.

http://www.phoronix.com/vr.php?view=24720

Nice, but what was the purpose of changing the test set?

[Michael]

Nice, but what was the purpose of changing the test set?

The test set was basically the same? There were just a few additions I added later on that were an oversight for not adding at the start of the tests.

[tildearrow]

What happened to Triangle? I want to see if the Fury still is at top, since the 1080 Ti hit 400000 a few days ago.

[fa5hion]

The test set was basically the same? There were just a few additions I added later on that were an oversight for not adding at the start of the tests.

And you removed one test where Nvidia would have gotten spanked. Is that test not reliable because it doesn't work well with Nvidia or is there similar behavior with AMD cards?

[fa5hion]

What happened to Triangle? I want to see if the Fury still is at top, since the 1080 Ti hit 400000 a few days ago.

That would be two removed tests then. Although the triangle test doesn't tell much about real world performance.

[edoantonioco]

Soon linux having slow GPU drivers will be a thing from the past.

[Jabberwocky]

Thanks for doing these tests! It's nice to have a RX 580 on the list, now I have a better indicator for my overclocked RX 480.

Are there any news about games or game technology that supports multi-gpu rendering (not talking about SLI or CrossFire)?

[leipero]

There's currently a problem with shaders on r600 running GpuTest 0.7.0 PixMark (Piano etc.) benchs on stable Mesa 17.1.0, forcing GL/GLSL 4.1 would run most of 4.0 benchmarks except PixMark FP64 and those with shader problem, there's already bug report on it (it seems to be old problem). Hopefully that soft-fp64 will get implemented you've mentioned in article from few days ago.