[Phoronix] 51 GPUs benched! Perf/W, regression, etc.

[nathanddrews]

This is some good stuff, too bad it's Linux only...

https://www.phoronix.com/scan.php?page=article&item=51-phoronix-13&num=1

AMD:
- FirePro V4900
- FirePro V8750
- Radeon HD 2900XT
- Radeon HD 3850
- Radeon HD 4670
- Radeon HD 4770
- Radeon HD 4850
- Radeon HD 4870
- Radeon HD 4890
- Radeon HD 5830
- Radeon HD 6870
- Radeon HD 6950
- Radeon HD 7950
- Radeon R7 260X
- Radeon R9 270X
- Radeon R9 285
- Radeon R9 290
- Radeon R7 370
- Radeon RX 460
- Radeon RX 480
- Radeon RX 550
- Radeon RX 560
- Radeon RX 580
- Radeon R9 Fury

NVIDIA:
- GeForce 8600GTS
- GeForce 8800GT
- GeForce 9600GSO
- GeForce 9800GT
- GeForce 9800GTX
- GeForce GT 240
- GeForce GTX 460
- GeForce GT 520
- GeForce GTX 550 Ti
- GeForce GT 610
- GeForce GTX 650
- GeForce GTX 680
- GeForce GTX 750 Ti
- GeForce GTX 760
- GeForce GTX 780 Ti
- GeForce GTX 950
- GeForce GTX 960
- GeForce GTX 970
- GeForce GTX 980
- GeForce GTX 980 Ti
- GeForce GT 1030
- GeForce GTX 1050
- GeForce GTX 1050 Ti
- GeForce GTX 1060
- GeForce GTX 1070
- GeForce GTX 1080
- GeForce GTX 1080 Ti

 

[Bacon1]

Serious Sam is the only test that looks correct.

DOTA 2 a 1060 is 13 fps faster than a 1080 Ti / 980 Ti. 1070 and 1080 are even slightly faster than the 1060.

Same deal with Metro last light where a 970 is 16 fps faster than 1070.

Even if the game was 100% cpu bound, the GPUs should be even in perf not one way faster.

 

[[DHT]Osiris]

Serious Sam is the only test that looks correct.

DOTA 2 a 1060 is 13 fps faster than a 1080 Ti / 980 Ti. 1070 and 1080 are even slightly faster than the 1060.

Same deal with Metro last light where a 970 is 16 fps faster than 1070.

Even if the game was 100% cpu bound, the GPUs should be even in perf not one way faster.

Yeah, there's a few things on that I feel like I wanna call shens on. The temp monitors look wonky too, like the 9800GTX capping out at 60c? I had an 8800GTX that could heat a 500sq ft dorm room in a german winter (with a window open!) like a champ.

Something tells me running all this on *nix threw the numbers.

 

[nathanddrews]

I doubt there are any shenanigans. There are known problems with the state of drivers and support on Linux, so it's far from a perfect test across the board, but there aren't many real world (non synthetic) tests one can run across so many GPUs using a single API (OGL). Half those GPUs don't support Vulkan...

Let's face it, this is partly the reason Linux isn't the de facto PC gaming OS.

 

[richaron]

Let's face it, this is partly the reason Linux isn't the de facto PC gaming OS.

I suspect Linux support is more a chicken/egg feedback kind of situation. But I don't argue with your point. Linux support for little stuff like temp sensors is lacking, and of course drivers themselves are fragmented and a lesser priority for the big companies.

That being said I'm still a happy Linux gamer with "good enough" performance and more than enough gaming content for me to waste my life away... Articles like this are a mammoth effort. So whilst it's impossible to have an all encompassing comparison of cards and drivers, this does provide an informative snapshot.

 

[Face2Face]

Yeah, there's a few things on that I feel like I wanna call shens on. The temp monitors look wonky too, like the 9800GTX capping out at 60c? I had an 8800GTX that could heat a 500sq ft dorm room in a german winter (with a window open!) like a champ.

Something tells me running all this on *nix threw the numbers.

With a TDP of 155w? Must of been gaming for quite some time

 

[poofyhairguy]

Good stuff. I have put together a little Linux machine just to play CS:Go at a desktop and I have wondering if the 750 ti I have or the 370 I have was the better choice. Looks like the answer is the 750 ti.

 

[[DHT]Osiris]

With a TDP of 155w? Must of been gaming for quite some time

I was! though to be fair, that could have had more to do with the intel q6800 than the 8800gtx.

 

[TeknoBug]

GTX950 better than GTX1050Ti? That's a bit weird.