NVIDIA GeForce GTX 980 Ti: Simply The Best For Linux Gamers

[phoronix]

Phoronix: NVIDIA GeForce GTX 980 Ti: Simply The Best For Linux Gamers

For the past few weeks I've been extensively testing the NVIDIA GeForce GTX 980 Ti on Linux and it's been a rather pleasant experience. Compared to the troubles with the R9 Fury on Catalyst Linux, the GTX 980 Ti has been a pleasant experience and yielding terrific results, assuming you're okay with using NVIDIA's proprietary driver.

http://www.phoronix.com/vr.php?view=21993

[dungeon]

What is the reason for this?

pts/csgo-1.1.0 [19 Jul 2015 14:25:50 EDT]
- Set enable threaded NV optimizations

I mean if you made that force working for benchmarking or something like that, then renaming CS:GO binary to profile with perf gain for AMD is also valid to me:

http://www.phoronix.com/scan.php?pag...rkaround&num=3

Why puting variabile in pts/csgo, just after you found that AMD Fury will for sure win in that benchmark

[Nobu]

Verses...verses...verses...verses--it's a slobberknocker!!!

[Michael]

What is the reason for this?

I mean if you made that force working for benchmarking or something like that, then renaming CS:GO binary to profile with perf gain for AMD is also valid to me:

http://www.phoronix.com/scan.php?pag...rkaround&num=3

Why puting variabile in pts/csgo, just after you found that AMD will win in that benchmark

It was pointed out in another forum thread (you can search for it, I think you even pointed it out or might have been Kano or someone) that it's present within Steam's launch scripts for the game but was missing from my PTS wrapper script of the binary. So it's just to match the behavior of Steam's package, as opposed to doing something else. Binary names are the same as the Steam package.

[Nobu]

As far as Per-per-watt, it looks like there are better options than the 980 Ti, especially in the nvidia camp. Even the Radeon R9 Fury only has an 11% deficite in that regard, and uses less power overall, so if you're on a power-constrained system it might be a better option. Of course, if you're willing to spend that much, you might as well upgrade your PSU too...

[dungeon]

And where is renamed binary for CS:GO for Fury benchmark, i want to see how 980Ti and Titan X lose from cheaper Fury, that is what should happen in that case.

That is where you should have also more proper power usage, perf per watt, etc... results.

[heliosh]

I would appreciate some OpenCL benchmarks, particularly Fury X vs. 980 Ti.

[Michael]

I would appreciate some OpenCL benchmarks, particularly Fury X vs. 980 Ti.

If you click on the Fury (I have no X) review and other articles, there are such benchmarks.

[torsionbar28]

I would appreciate some OpenCL benchmarks, particularly Fury X vs. 980 Ti.

Why bother? 980 ti is crippled in double precision floating point. Right now, the GK110 based Titan (and Titan Black and Titan Z) are the only two consumer nvidia cards with full double precision (FMA) performance.

All 900 series, including Titan X, are crippled.

Aside from original Titan, only Quadro K6000 or Tesla series have full performance FMA. I guess if you don't need FMA, it's a non issue, but it's hard to ignore FMA in a discussion about GPU computing...

[ElectricPrism]

$630 - $770 for the best Linux Gaming Experience minus other hardware. What a chunk of change, wow.