AMD & NVIDIA GPU VR Performance: Project Cars @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
54,911
AMD & NVIDIA GPU VR Performance: Project Cars - If you like racing games then it is likely that you are familiar with Project Cars. This racing game can push a system to its limits when you turn up the visuals, and that is just what we are doing today while using our HTC Vive VR headset. And this is certainly a game we will add to the VR Leaderboard.
 
So, Assetto Corsa is too unbiased to test?
 
Last edited:
Remon said:
So, Assetto Corsa is too unbiased to test?
Click to expand...

Doesn't Assetto Corsa only support Oculus at this time? Officially at least since LibreVR Revive is needed for HTC Vive support.

And if you want a test for Assetto Corsa, why not simply ask if one is planned instead of making snide remark about a supposedly biased test?
 
Kyle has a Titan X. That's all I got from this. lol
Waiting on review.

But I am surprised the 1080 and 1070/980Ti are all about the same in this game.
I was thinking the 1080 was going to be in the single digits as well.
 
I bought a GTX 1080 and then ordered but returned the Titan X Pascal. After reading this along with my own experience in Raw Data - I'm wishing I kept it!

Hoping VR SLI support is around the corner, I think that would be the better compromise, 2xGTX 1080's > 1xTitan for roughly the same price.
 
You must be a mind reader. You have reviewed every game I was going to suggest before I ever did.

BTW why did you do the testing with an outside view instead of cockpit view as that's the one people would normally use?
 
Have to say I'm really glad for this particular test.

I'm really thinking that simulation games is the big field for VR to occupy because we no longer need to build multi monitor space wasting setups to get all around views of your cockpit. The VR headset is your all around view.

However, this test also suggests to me that only the Titan X can play this game. Without knowing what constant reprojection looks like, I can only surmise it's bad. It's one thing to describe reprojection in text but I just need to see and experience it for myself.

Kyle, care to do a review with this? FlyInside: Oculus Rift Support for FSX & Prepar3D

Edit: isn't this sorta....F35 pilot's helmet?
 
Last edited:
VitaminTee said:
I bought a GTX 1080 and then ordered but returned the Titan X Pascal. After reading this along with my own experience in Raw Data - I'm wishing I kept it!

Hoping VR SLI support is around the corner, I think that would be the better compromise, 2xGTX 1080's > 1xTitan for roughly the same price.
Click to expand...

VR SLI doesn't scale well (~30%) because you have to do split frame rendering rather than alternate frame rendering. Combined with single pass stereo which can increase single card performance by 20-30%, I think you will see a single titan xp basically matching 1080 sli in VR.

And that is assuming widespread SLI VR support ever materializes.

ChrisC said:
Have to say I'm really glad for this particular test.

I'm really thinking that simulation games is the big field for VR to occupy because we no longer need to build multi monitor space wasting setups to get all around views of your cockpit. The VR headset is your all around view.

However, this test also suggests to me that only the Titan X can play this game. Without knowing what constant reprojection looks like, I can only surmise it's bad. It's one thing to describe reprojection in text but I just need to see and experience it for myself.
Click to expand...

It's a PC game with a ton of graphics settings. There's plenty of stuff you can drop down to make it run well on lesser cards.
 
Daniel_Chang said:
Project CARS is still one of the top 5 played games on Steam. Let that sink in for a moment. It has nothing to do with bias, and everything to do with relevance.
Click to expand...

Is it one of the most played VR games? Is it played more in VR than assetto corsa?

If player count of the base game is the criteria then minecraft VR and dota 2 VR should be ahead of project cars


Also I'm not sure where the hell you're getting your info from. I don't see project cars on the list of the 100 most played games on steam. There are 955 people playing it right now so it probably wouldn't even make the top 500. Dota 2 had a peak today of 1.028 million. 892k currently.

Assetto Corsa has 1500 players right now btw


Edit: Furthermore, Project Cars' all time players peak is 10k, which wouldn't even put it in the top 25 today on steam. Are you thinking of a different game?
 
Last edited:
Stoly said:
Whats the deal with Assetto Corsa anyway?
Click to expand...

It's a driving sim, while PCars is a simcade. For some "weird" reason, even though Assetto Corsa has better physics, they aren't as taxing as Pcars' own.
 
Tried the exact settings with 5820k overclocked to 4.4 and 980ti overclocked. CV1 performance indicator showed latency 28%, never below 90 fps and 2 dropped frames for the lap. Used in car view and was turning head constantly. No judder and for me no motion sickness. Almost all racings sims at this time seem to work better on Nvidia. Most are relatively low budget so less money to spend on multi gpu manufacturers optimization.
 
pj- said:
Is it one of the most played VR games? Is it played more in VR than assetto corsa?
Click to expand...

That's a fair question, see below.

pj- said:
If player count of the base game is the criteria then minecraft VR and dota 2 VR should be ahead of project cars
Click to expand...

A game's relevant has more to do with than JUST player count. Kyle and his staff can speak better to this, but here's my criteria:

Age - Newer titles are USUALLY more relevant than older titles
Current players - This will contradict age sometimes, if a game is still popular then it's worth investigating/reviewing.
System demands - CS Go has MANY active players, but it doesn't push a system so it doesn't matter in most modern reviews.

When you add up those factors, and many others I'm sure, you can determine which games are worth testing. And every reviewer has a different conclusion, so there should be no surprise that laymen like you and I disagree.

pj- said:
Also I'm not sure where the hell you're getting your info from. I don't see project cars on the list of the 100 most played games on steam. There are 955 people playing it right now so it probably wouldn't even make the top 500. Dota 2 had a peak today of 1.028 million. 892k currently.
Click to expand...

From the second paragraph in the article that this thread is about:
Players Graphs

I did misspeak, and I apologize for that. I should have specified top 5 in Steam VR. PCars ranks 4th. EDIT: 4th in current users at time of post, but in a near dead heat with 5th, so they could swap.

pj- said:
Edit: Furthermore, Project Cars' all time players peak is 10k, which wouldn't even put it in the top 25 today on steam. Are you thinking of a different game?
Click to expand...

When I made my post, I mischaracterized the way that the article wrote it. That was my mistake, see above.
 
There were some whiny fucks complaining that [H] was only benchmarking games based on Unreal. Well, you have your game now. Happy?
 
Looks like unless I win the lottery, I won't be playing VR games anytime soon. The RX 480 is reasonably priced, but... the $1000.00 headset... holy crap, yeah, I have other bills to pay first.
 
Any chance of seeing the same test but with the cards overclocked? Almost every Pascal card makes it to 2GHz easy, and the 1080 might be able to cut down on reprojection mode quite a bit given a little boost.
 
jihadjoe said:
Any chance of seeing the same test but with the cards overclocked? Almost every Pascal card makes it to 2GHz easy, and the 1080 might be able to cut down on reprojection mode quite a bit given a little boost.
Click to expand...
Nope. Moving on to the next game now.
 
Daniel_Chang said:
That's a fair question, see below.



A game's relevant has more to do with than JUST player count. Kyle and his staff can speak better to this, but here's my criteria:

Age - Newer titles are USUALLY more relevant than older titles
Current players - This will contradict age sometimes, if a game is still popular then it's worth investigating/reviewing.
System demands - CS Go has MANY active players, but it doesn't push a system so it doesn't matter in most modern reviews.

When you add up those factors, and many others I'm sure, you can determine which games are worth testing. And every reviewer has a different conclusion, so there should be no surprise that laymen like you and I disagree.



From the second paragraph in the article that this thread is about:
Players Graphs

I did misspeak, and I apologize for that. I should have specified top 5 in Steam VR. PCars ranks 4th. EDIT: 4th in current users at time of post, but in a near dead heat with 5th, so they could swap.



When I made my post, I mischaracterized the way that the article wrote it. That was my mistake, see above.
Click to expand...

Ah, I didn't know anyone was tracking top games with VR support. Assetto Corsa wouldn't appear on there because it only currently supports oculus.

Still, it doesn't say how many of those people are playing in VR. 1000 people are playing project cars, but what if only 8 are using VR? Looking at the rest of those player counts, it is not out of the realm of possibility.

I think it's a fine game to test because there haven't been any non-vr native games so far and it can be configured to be pretty demanding. Elite Dangerous probably would have been better to do first since I've seen way more people talk about that in VR than project cars. It is similar in age, more popular, and I believe it also can be hard on GPUs.

I'd also like to see some ReVive testing just out of curiosity
 
I believe that @Kyle should rank the FuryX in front of RX480. FuryX's high price can't outweight RX480's tragic VR performance in my opinion.
(all the other GPUs presented <1% dropped frames, while RX480 presented >18% dropped frames:eek: )
(P.S. Another VR victory for NVidia, 5 out of 5 so far.!!:p )
 
Evil said:
Looks like unless I win the lottery, I won't be playing VR games anytime soon. The RX 480 is reasonably priced, but... the $1000.00 headset... holy crap, yeah, I have other bills to pay first.
Click to expand...

1) Vive is $800, Rift is $600

2) I wouldn't buy a RX480 for a VR rig in the first place.
 
On the final page it say avg render time for fury x is 11.7ms, yet when i look at the graph for the card it looks like it should be closer to 19-20ms. same can be said for rx 480 and gtx1060 as the render times seem least to me on last page don't match the graph.

Daniel_Chang said:
Project CARS is still one of the top 5 played games on Steam. Let that sink in for a moment. It has nothing to do with bias, and everything to do with relevance.
Click to expand...
Same people complaining about biased are same people would claim using AOTS is a valid game to test with even with how biased it is towards AMD cards and even ignoring fact only ~60k people even own it.
 
Arbit3r said:
On the final page it say avg render time for fury x is 11.7ms, yet when i look at the graph for the card it looks like it should be closer to 19-20ms. same can be said for rx 480 and gtx1060 as the render times seem least to me on last page don't match the graph.
Click to expand...
You are looking at the [H]ardOCP VR Leaderboard data which is comprised of data averaged from Robot Repair, Raw Data, Call of Starseed, and Project Cars. Just like it states at the top of the Leaderboard. :)

1472083550cURQD9x1c8_6_2.png
 
Sith'ari said:
I believe that @Kyle should rank the FuryX in front of RX480. FuryX's high price can't outweight RX480's tragic VR performance in my opinion.
(all the other GPUs presented <1% dropped frames, while RX480 presented >18% dropped frames:eek: )
(P.S. Another VR victory for NVidia, 5 out of 5 so far.!!:p )
Click to expand...

I think it is fine the way it is, considering the Fury X still cost around $400, for that price, you are better off buying a 1070 or 980Ti. At least with the 480, it has price going for it.
 
Wait, did I read that there was no performance penalty to going to the higher AA setting from Zero/None? Or is that only with everything set to low/off?
 
its funny cause i cant imagine dota2 being ANYTHING special in VR? For real lol....Those others yea...the driving/flying games i can imagine being amazing! whats so special about dota2 lol
 
I love my 290X's but man AMD is getting spanked on the VR front...hopefully they can fix the drivers to unleash the true potential of their chips. I want competition, not a blow out. Is CF or SLI working for any VR titles? I want to pick up the next iteration of the Vive and had hoped my 290X's would give me a year or so of GPU horsepower
 
AMD having trouble with high FPS games such as VR games, who would have guessed. I don't understand the AotS criticism though guys, AotS isn't popular but it's a decent compute benchmark, some people even say there's a game mode hidden in it too.

PCars is savage on the amd front, probably a little more savage than other games.

I honestly don't see AMD being competitive in VR, when have AMD products ever handled high-fps gameplay well? The only game I can think of is doom in vulkan and that a very flat game for the most with not much detail
 
Chimpee said:
Funny you said that, even AMD own website said VR Ready Premium Products.
Click to expand...
How many AMD buyers can afford the VR stuff yet anyway? Some, but Not many.....We will start to get very concerned in a couple of years! Until then not so much! I will agree, so far in VR its not looking so pretty
 
primetime said:
How many AMD buyers can afford the VR stuff yet anyway? Some, but Not many.....We will start to get very concerned in a couple of years! Until then not so much! I will agree, so far in VR its not looking so pretty
Click to expand...

Hand-waving the cost factor (which has been brought already by unbiased observers in various threads) does not negate the fact that AMD markets this as VR for the masses and a premium VR experience.

It seems that only one company took their PR VR claims seriously.
 
Algrim said:
Hand-waving the cost factor (which has been brought already by unbiased observers in various threads) does not negate the fact that AMD markets this as VR for the masses and a premium VR experience.

It seems that only one company took their PR VR claims seriously.
Click to expand...
And now it seems Vega is pushed out at least 6 months from now.

amd-investor-presentation-2016_Page_08.jpg
 
Most of the rumors have been pegging Vega at first quarter of 2017 whereas this slide distinctly targets first half. (Yes, there was that one poster that claimed October 2016 but most sane predictions were beginning of 2017.)
 
Enthusiast Vega is Vega 11.

Vega 11 is 1H 2017.

Vega 10 should at least launch this year.

That said if Vega 11 launches next summer it's going to be hilariously late, NV will have had a year of Titan X sales and will probably counter with a ~400 x70 Volta part with similar performance.

Volta next summer seems pretty likely Imo
 
You must log in or register to reply here.
Back
Top