SteamVR Performance Test: 16 GPUs Compared

[kcarbotte]

The score for the 295x2 is wrong, there is a manual setting for in SteamVR that enables proper Crossfire use, my 295x2 with a 2600K @ 4.8ghz gets a score of 9.7, after changing the setting, before making the change I get the score listed above.

I concur with Sergio_9. Here is my 295x2 score http://imgur.com/6b7QiO9. Article is dead wrong! Please fix!

The score that we got isn't wrong, but its entirely possible there is something wrong with CrossFire on this motherboard.
We ran the test on another system (X99 5930K) and the score was something like 9.6.

As I stated, the test was reran more than once. There was definitely something wrong with that config, but the settings are correct.

Also, as per AMD, you do not have to add the multigpu tag to the startup of the app. The SteamVR Performance Test is fully compatible with CrossFire. You need to use that tag for SLI though, because the Nvidia drivers aren't optimized to use two GPUs for VR yet.

 

[CoNk3r]

Guys, you discarded de gtx 950 at the beginning, but it result to have a stable frame rate close to the 970...(asus strix). being realistic about having a budget machine, how did you feel the experience with VR over this GPU?

 

[rainyfeels]

Hj

Is some of this based off architecture? My 780 is superclocked and can easily compete with a 970 yet only gets 4.5 in the test and is "capable". Or will it be fine when it actually comes to playing games?

That's a great question.
I didn't have a GTX 780 to test, so I couldn't test that out.
I was very curious myself about the score, but your CPU will play a roll in the overall score. If you're running an AMD CPU it may not have the single thread computation power to keep up.

It's also importatnt to keep in mind that your GTX 780, while comparable to a 970, does not actually keep up with a 970 in all tests.
According to Nvidia's own graph, even the Titan is bested by the 970 in 3DMark FireStrike.
http://international.download.nvidia.com/webassets/en_US/shared/images/products/shared/lineup-full.png

My CPU is a i7 2600k so I know it won't be that, hard to think that its only just better than a R9 380...Must just be architecture and drivers etc. I would like to know how my 780 would perform though in a real game as I do have a Vive pre ordered at the moment and am waiting for the next gen of cards.

 

[sephirotic]

I can't take serious any minimum recommendation that puts a haswell as cpu recommended. I've yet to see a single game that would be hurt by the 20~30% less performance per clock than my good old 2500k running at 4.4ghz.
I run the SteamVR test, and I got a 8.3 score, on my 2500k with an overclocked 970 at 1526mhz. Frames bellow 90 were 0, cpu bond frames were only 5 (0%), with fidlity score around Very high.

You don't need a 4th generation Intel processor for VR. All you need is a sandy bridge.

 

[kcarbotte]

Guys, you discarded de gtx 950 at the beginning, but it result to have a stable frame rate close to the 970...(asus strix). being realistic about having a budget machine, how did you feel the experience with VR over this GPU?

It had stable framerates, but the image quality was very low. Valve made the SteamVR Performance Test with the Source 2 engine and is able to scale the details down accordingly.
Most games aren't built with this engine, and likely won't scale down as well. As the test says, it's recomended that you have a stronger GPU for VR.

I plan to test the experience first hand in the near future though. Once we have some final builds of games, I'll be trying them out with a range of cards to see how real games perform.
Watch for those tests once the consumer hardware is finally available.

 

[rhysiam]

Anyone care to comment about why we're seeing basically no gains whatsoever from multi-GPU configs in the fidelity scores? The forums are full of people claiming that VR will finally start to make proper use of two card configs (I've probably made similar claims myself).

I realise it's early days, but seeing a 390 match a 295X2, and (even worse) a single OC'd 950 beat SLI'd 960s is concerning is it not? Are these just driver issues we're seeing?

 

[OriginFree]

10.8 on a 980M with a Intel i7-4710HQ @2.50GHz (8 cores) and that was a few drivers ago.

To me this would indicate that the GPU is the deciding factor once you hit some predetermined "must be better than X" CPU level. I am curious if mobile graphics have issues vs desktop (frame control etc) and how accurate the scaling is for the GPU. Given the power of the GPUs that scored 11, mine should be nowhere near them.

Hey, tom's. When you get a chance maybe test out a few laptops (gaming and not so) to see if VR on the go is a possibility. I would love to see what the minimum laptop spec would be for VR parties, etc. Hmmm and maybe some desktop GPUs that would work with SFF systems.

Nothing like a bunch of nerds in a living room with VR headsets talking in a virtual world instead of the real one, as a comment on our technological social skills. Well at least there's no VR facebook / twitter (yet).

 

[Lavmatic]

The recommended PC specs from htcvive.com lists the AMD GPU as the R9 290, not the R9 390 like called out in the article. I wonder how the R9 290 compares.

 

[mavikt]

So either you're VR ready, or your next GPU will be a Pascal or Polaris.
The GPU die-shrink is so long overdue and when they come there will be VR to the people. (hopefully!)

 

[InvalidError]

when they come there will be VR to the people. (hopefully!)

I predict that most people will get motion sickness and vertigo from VR due to stutter, tracking errors, lag, sensory contradictions/disorientation and other issues, then give up on it for most of the next decade, much like what happened with attempts at 3D glasses in the past and 3D TV/LCD more recently. It may sound nice in theory and work in practice for some people, but might not be ripe for the mainstream until much later.

 

[Sergio_9]

The score for the 295x2 is wrong, there is a manual setting in SteamVR that enables proper Crossfire use, my 295x2 with a 2600K @ 4.8ghz gets a score of 9.7, after changing the setting, before making the change I got the score listed above.

I believe the setting is the -vrforce set in Steam command line setting for the benchmark. I don't have access to my PC at the moment to confirm. The Steam forum is where you can find more details. Also, for reference, the AMD driver version I ran this benchmark was Crimson 16.2.1, on Windows 10 Pro, current regular build.

 

[TechyInAZ]

when they come there will be VR to the people. (hopefully!)

I predict that most people will get motion sickness and vertigo from VR due to stutter, tracking errors, lag, sensory contradictions/disorientation and other issues, then give up on it for most of the next decade, much like what happened with attempts at 3D glasses in the past and 3D TV/LCD more recently. It may sound nice in theory and work in practice for some people, but might not be ripe for the mainstream until much later.

I will agree that it is probably not something that will takeoff. But, with the bug fixes and advantages of DX12, we could have less stuttering issues than we might expect in the upcoming years.

Then again, I haven't tried it yet so I don't know much.

 

[Sergio_9]

The score for the 295x2 is wrong, there is a manual setting in SteamVR that enables proper Crossfire use, my 295x2 with a 2600K @ 4.8ghz gets a score of 9.7, after changing the setting, before making the change I got the score listed above.

I believe the setting is the -vrforce set in Steam command line setting for the benchmark. I don't have access to my PC at the moment to confirm. The Steam forum is where you can find more details. Also, for reference, the AMD driver version I ran this benchmark was Crimson 16.2.1, on Windows 10 Pro, current regular build.

The proper way to configure SteamVR is in the launch options, in Steam, to enter -multigpu, this enables crossfire. The crossfire setting in the Crimson driver can stay at default, no need to change it.

 

[mamasan2000]

Is some of this based off architecture? My 780 is superclocked and can easily compete with a 970 yet only gets 4.5 in the test and is "capable". Or will it be fine when it actually comes to playing games?

That's a great question.
I didn't have a GTX 780 to test, so I couldn't test that out.
I was very curious myself about the score, but your CPU will play a roll in the overall score. If you're running an AMD CPU it may not have the single thread computation power to keep up.

It's also importatnt to keep in mind that your GTX 780, while comparable to a 970, does not actually keep up with a 970 in all tests.
According to Nvidia's own graph, even the Titan is bested by the 970 in 3DMark FireStrike.
http://international.download.nvidia.com/webassets/en_US/shared/images/products/shared/lineup-full.png

I wouldn't say AMD CPU can't keep up. I have FX-8350 @ 4.8 Ghz and Radeon 290X, scored a 7.3 score on the steamVRtest. Not a single frame under 90 fps.
Better than in i5 with a GTX 970.

 

[Fa1l3d]

ARE U KIDDING ME ????? I have 2.1 points on this test whit my GTX770/8Gb/I7-3770S.

 

[cats_Paw]

So... Expensive VR, that needs expensive GPUs to run well, with almost no games out there for it yet...
Em... wait a bit?

 

[MackLuster]

My Sapphire R295X2 scored an 11, so....

 

[Marcus_Vraeden]

Ran it with a 980ti, AMD Phenom II 965 (3.4ghz, 4 core) which I believe is below rec. spec, and 8 GB of RAM, scored an 11. Very interested to see real game performance with this old CPU.

 

[picture_perfect]

I got a 8.3 score, on my 2500k with an overclocked 970 at 1526mhz.
You don't need a 4th generation Intel processor for VR. All you need is a sandy bridge.

Hopefully then Sandybridge owners like us would just add a gtx980 + power supply and be done with it. SandyBridge is limited to pcie 2.0, but if anyone is wondering, that makes no difference.:

http://www.techpowerup.com/reviews/NVIDIA/GTX_980_PCI-Express_Scaling/22.html

 

[LORD_ORION]

You're going to be very angry spending $1000 on gear that was capable of doing the same thing 20 years ago using magic carpet... only to find that after an hour 75% of you will have a migraine, and after 2 hours 99% of you will have temporary vertigo.

 

[InvalidError]

You're going to be very angry spending $1000 on gear that was capable of doing the same thing 20 years ago using magic carpet... only to find that after an hour 75% of you will have a migraine, and after 2 hours 99% of you will have temporary vertigo.

That's exactly what I am thinking too. Playing 3D games on a regular 2D screen already makes me a little dizzy sometimes after an hour or two and I cannot play 3DS games with 3D enabled for more than 10 minutes at a time when I tried one of my friends'. It will probably take 10-20 years before VR gets perfected enough to eliminate those side-effects.

 

[envain]

Because I have an i7-2600k (running at 4.8Ghz), I apparently don't meet the minimum requirements for Oculus Rift. Really?

They don't really bother checking your CPU. If it's older gen, they automatically assume it sucks. funny thing is that I also run a 2600k @ 4.8Ghz and even systemrequirementslab says my CPU can't run latest games (not in VR). Yet I play everything on Ultra...

I think it might be bugged, I passed with a 9.1 and I'm running my i7 870 at 3.5.

 

[Math Geek]

the real mainstream adoption won't happen until the consoles have it. then the prices will be finally low enough for people to really give it a chance. i have had many people ask me about it as a consideration with a new build but once the specs are priced, most of them shrug and say "maybe later". the ones who already have the specs in the build, balk at the price of the hardware.

 

[mavikt]

when they come there will be VR to the people. (hopefully!)

I predict that most people will get motion sickness and vertigo from VR ...

My observation is that more people are soon going to be eligible to run these new VR toys at their desired specs (the price of admission is going down).

If ppl have any medical condition that will prevent them from enjoying it I don't know. I know one person at work that does have, but the other (families) seem to enjoy 3D from bluray movies at least.
Any time the experience IRL differ from the VR one, there is source if 'misalignment'. Normally you're 'grounded' IRL, but if in a VR experience you're jumping and you're not jumping IRL, there's a misalignment in the experience which could provide discomfort.
I think as long as the VR experience matches your current situation you're ok. You'd need to feed the brain a believable picture!
Though, this rule would be so easy to break, and that would break to whole VR experience, unless they'd invent a way to get around it...

 

[cherie22984]

295x2 score is way off? Here is mine http://imgur.com/GhrMEfm