Intel Celeron/Pentium/Core i3/i5/i7 - NVIDIA vs. AMD Linux Gaming Performance

[Zarathustra[H]]

Those Open Source fanatics over at Phoronix have a nice comparison of how 5 different Intel CPU's with 5 different AMD and Nvidia GPU's perform in some of the most popular games available in Linux. They keep the commentary short and sweet, and mostly let their results do the talking, so if you like nice charts full of data pitting hardware against hardware, be sure to head on over there and check them out.

There don't appear to be any major surprises, other than maybe that the AMD GPU's seem to be performing slightly worse under Linux than they do under Windows, but it is still nice see a good comparison. Personally, I wish more titles would work well under Linux. I transitioned to Linux as my main operating system long ago, but still keep a Windows install around for my gaming needs.

If you've been weighing CPU and GPU options for a Linux gaming system upgrade, hopefully you found these multiple data points useful. Those wishing to see how their own GPU/CPU compare to the assortment of hardware used in this article, simply install the Phoronix Test Suite and run phoronix-test-suite benchmark 1702083-RI-KABYLAKEG39.

 

[Dekoth-E-]

People still buy Celeron?

 

[Zarathustra[H]]

People still buy Celeron?

I have two G1840's. Granted, not for gaming though. My three HTPC's (i5-4570T, and two G1840's) all have GeForce 720GT's and do nothing but serve up videos from my NAS, usually via MythTV.

The way I see it is, if no one tests them, how do we know they aren't good enough for the job?

 

[MMitch]

Actually surprised by the difference between i5 and i7, any thoughts ?

 

[Zarathustra[H]]

Actually surprised by the difference between i5 and i7, any thoughts ?

Looking at it, there is a ~10.5% difference in base clock and a 7.1% difference in max single core turbo clock.

The difference between i5 and i7 is about 13% on the 1080 in the first benchmark, so right off the bat, I think a good chunk of that is up to clock speed differences.

The little that remains might be the engines of these games taking advantage of hyperthreading.

It's not surprising that we'd see different scaling under Linux than under Windows due to the differences in how the schedulers work.

 

[MMitch]

Looking at it, there is a ~10.5% difference in base clock and a 7.1% difference in max single core turbo clock.

The difference between i5 and i7 is about 13% on the 1080 in the first benchmark, so right off the bat, I think a good chunk of that is up to clock speed differences.

The little that remains might be the engines of these games taking advantage of hyperthreading.

It's not surprising that we'd see different scaling under Linux than under Windows due to the differences in how the schedulers work.

Thanks for the detailed explanation, makes sense if they weren't clocked the same.
That brings another point, why not test those at the same core clocks ? Still a great comparison, not complaining.

 

[Zarathustra[H]]

Thanks for the detailed explanation, makes sense if they weren't clocked the same.
That brings another point, why not test those at the same core clocks ? Still a great comparison, not complaining.

That is something that Kyle has done in his new CPU reviews in the past, and probably (I can't speak for him) will do when new architectures are released. (like Ryzen for instance)

This review is not ours though, so we have no input as to what they tested.