ComputerBaseVirtual Reality Benchmarks

[Sweepr]

www.computerbase.de/2016-03/systemanforderungen-vr-spiele-benchmarks

 

[railven]

I forget which is which, Steam VR is the Nvidia joint project, correct? Also HTC Vive?

And Oculus is AMD Liquid VR?

 

[Sweepr]

Showcasing next-generation VR-optimized GPU hardware – AMD today demonstrated for the first time ever the company’s forthcoming Polaris 10 GPU running Valve’s Aperture Science Robot Repair demo powered by the HTC Vive Pre. The sample GPU features the recently announced Polaris GPU architecture designed for 14nm FinFET, optimized for DirectX® 12 and VR, and boasts significant architectural improvements over previous AMD architectures including HDR monitor support, industry-leading performance-per-watt2, and AMD’s 4th generation Graphics Core Next (GCN) architecture.

http://videocardz.com/58486/amd-to-demonstrate-polaris-10-gpu-in-steam-vr-benchmark-at-gdc2016

Perhaps we will find out Polaris 10 score later today.

 

[dark zero]

Seems that except the last one, every game are well optimized. The last one needs optimization in order to become decent.

 

[railven]

http://videocardz.com/58486/amd-to-demonstrate-polaris-10-gpu-in-steam-vr-benchmark-at-gdc2016

Perhaps we will find out Polaris 10 score later today.

Feels like Gemini is already dead.

 

[Kenmitch]

Feels like Gemini is already dead.

Gemini could be for the elitest market and little Polaris for the common man. Bringing entry level VR at a reasonable price(ignoring headset cost) is one of AMD's goals.

What's Starbucks without nerds on laptops with VR headsets.

 

[pj-]

I forget which is which, Steam VR is the Nvidia joint project, correct? Also HTC Vive?

And Oculus is AMD Liquid VR?

I don't think the partnerships you're describing exist. In the infancy of VR, the headset makers would be stupid to not have their systems perform as well as possible on whatever computer.

"The [steamvr] readiness test also implements one of AMD’s LiquidVR features known as Affinity multi-GPU, which lets each GPU render a scene for just one eye. AMD does note, however, that while multi-GPU is working in the SteamVR Performance Test, it’s nowhere near being finished, although “significant uplift” in performance is possible over a single GPU’s performance."

http://www.forbes.com/sites/jasonev...nce-test-no-vr-headset-required/#56a692307c2e

Conversely, all of the "oculus ready" PCs listed on oculus.com come with nvidia GPUs.

 

[Sweepr]

Feels like Gemini is already dead.

Too little too late. All the fun belongs to Pascal/Polaris GPUs now.

 

[Azix]

dx11 vr, not interesting

 

[railven]

I don't think the partnerships you're describing exist. In the infancy of VR, the headset makers would be stupid to not have their systems perform as well as possible on whatever computer.

"The [steamvr] readiness test also implements one of AMD’s LiquidVR features known as Affinity multi-GPU, which lets each GPU render a scene for just one eye. AMD does note, however, that while multi-GPU is working in the SteamVR Performance Test, it’s nowhere near being finished, although “significant uplift” in performance is possible over a single GPU’s performance."

http://www.forbes.com/sites/jasonev...nce-test-no-vr-headset-required/#56a692307c2e

Conversely, all of the "oculus ready" PCs listed on oculus.com come with nvidia GPUs.

I'm just going about what I recall from around here. But I could be mistaken. And in another thread I just saw that some how AMD owns 83% of the market...

More to the issue, I personally was surprised to see NV leading in those benchmarks, especially after all the Liquid VR and AMD is dominating NV posts.

EDIT: Post above mine might be what I was missing.

Gemini could be for the elitest market and little Polaris for the common man. Bringing entry level VR at a reasonable price(ignoring headset cost) is one of AMD's goals.

What's Starbucks without nerds on laptops with VR headsets.

Too little too late. All the fun belongs to Pascal/Polaris GPUs now.

Gemini just seems to have been put on the back burner with Polaris looming. We still haven't seen official product info on the thing. I just don't seem AMD promoting Gemini when they can sell you Polaris.

 

[Jaydip]

Tom's test
http://www.tomshardware.com/reviews/steamvr-performance-test-gpu-comparison,4489.html

 

[PhonakV30]

AMD and Nvidia are opposite.

Our FRAPS data also suggests that AMD's cards tend to have more regulated frame rates. You can see from the minimum and maximum frame rate graphs that AMD GPUs target the minimum frame rate more accurately at 100 FPS across the board. In comparison, Nvidia seems to favor a higher minimum frame rate. But this data doesn't tell the full story.

Just look at Fraps Images
http://www.tomshardware.com/reviews/steamvr-performance-test-gpu-comparison,4489-2.html

 

[Bacon1]

SteamVR uses LiquidVR for AMD cards so it makes sense that its optimized well.

AMD is pleased to report that our Radeon R9 390, Nano, and Fury series GPUs are all able to achieve ‘VR Recommended’ status. This performance is a great testament of the work AMD has done with Valve and other technology partners to deliver the best VR experience with our products through the AMD LiquidVR™ initiative.

https://community.amd.com/community...-recommended-for-vr-by-steam-performance-test

Also why CFX is supported

 

[Headfoot]

I forget which is which, Steam VR is the Nvidia joint project, correct? Also HTC Vive?

And Oculus is AMD Liquid VR?

No, this is not yet the case.

There is a lot of cross marketing but there are no lines in the sand yet. Everyone is trying to support everyone at this point.

 

[Bacon1]

Did they mention what clockspeeds the 980 TI was running at? Its scoring a perfect 11 even though most people I've seen post results are mid 10s at best.

 

[tential]

Did they mention what clockspeeds the 980 TI was running at? Its scoring a perfect 11 even though most people I've seen post results are mid 10s at best.

I've seen perfect 11s... I

 

[pj-]

SteamVR uses LiquidVR for AMD cards so it makes sense that its optimized well.



https://community.amd.com/community...-recommended-for-vr-by-steam-performance-test

Also why CFX is supported

CFX is "supported" yet gives no zero benefit over a single card (as per Tomshardware's testing).

Did they mention what clockspeeds the 980 TI was running at? Its scoring a perfect 11 even though most people I've seen post results are mid 10s at best.

Only a reference 980ti should score below 11. Any factory OC model will max it out from what I've seen.

This is what a completely stock 980ti's results look like:

It doesn't take much to make that a completely flat line

 

[Bacon1]

CFX is "supported" yet gives no zero benefit over a single card (as per Tomshardware's testing).




Only a reference 980ti should score below 11. Any factory OC model will max it out from what I've seen.

This is what a completely stock 980ti's results look like:

It doesn't take much to make that a completely flat line

They probably didn't use the -multigpu launch command for it to put load on both GPUs.

295x2

And yes I know it doesn't take much, just wondering what they used. Having that information makes a difference in FPS testing.

 

[3DVagabond]

I don't think the partnerships you're describing exist. In the infancy of VR, the headset makers would be stupid to not have their systems perform as well as possible on whatever computer.

"The [steamvr] readiness test also implements one of AMD’s LiquidVR features known as Affinity multi-GPU, which lets each GPU render a scene for just one eye. AMD does note, however, that while multi-GPU is working in the SteamVR Performance Test, it’s nowhere near being finished, although “significant uplift” in performance is possible over a single GPU’s performance."

http://www.forbes.com/sites/jasonev...nce-test-no-vr-headset-required/#56a692307c2e

Conversely, all of the "oculus ready" PCs listed on oculus.com come with nvidia GPUs.

Got a source that's not behind an ad wall?

 

[3DVagabond]

Did they mention what clockspeeds the 980 TI was running at? Its scoring a perfect 11 even though most people I've seen post results are mid 10s at best.

I've seen mostly perfect scores for the 980 ti. We haven't seen the 980 ti consistently beat FuryX in games recently unless it's O/C'd. We'll have to see how that translates to real world gaming.

CFX is "supported" yet gives no zero benefit over a single card (as per Tomshardware's testing).




Only a reference 980ti should score below 11. Any factory OC model will max it out from what I've seen.

This is what a completely stock 980ti's results look like:

It doesn't take much to make that a completely flat line

Well, after seeing this, maybe all of those 11's I've been seeing are with O/C'd cards without it being revealed. Interesting. Thanks for posting this.

 

[pj-]

Got a source that's not behind an ad wall?

https://www.linkedin.com/pulse/valves-vr-ready-aperture-test-roy-taylor?trk=prof-post

 

[3DVagabond]

https://www.linkedin.com/pulse/valves-vr-ready-aperture-test-roy-taylor?trk=prof-post

Cheers! :thumbsup:

 

[Sweepr]

Tom's test
http://www.tomshardware.com/reviews/steamvr-performance-test-gpu-comparison,4489.html

 

[Silverforce11]

These raw frame-rate comparisons is akin to a VR experience where you stare at a scene static.

They don't measure the most important metric: motion to photon latency.

Tom's did a test last year on 3dMark's latency test (DX11) and it was awful, every GPU across the field got 35-50ms lag time from when you move your head, to when the scene is visually updated.

We still lack verifiable data and benchmarks for this latency that utilizes LiquidVR and GameWorks VR APIs.

 

[Bacon1]

Yeah these numbers look all wrong. Fury should have at least mid 9, and CFX is showing no improvements.

I even showed someone running their own 295x2 and it gave solid 11, unlike the 7.10 shown here.