[computerbase] Project CARS benchmarks

Page 14 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Eymar

Golden Member
Aug 30, 2001
1,646
14
91
I'm not sure anything can be done to convince those who do not believe words straight from the developer...

Yeah, the game obviously has a lot of work put into it (in my opinion best looking PC racing game) so must be disheartening to devs to see it being trashed (warranted or not). I tried the game on a 290x (14.12 drivers) with i7 920 (no overclock) and can get decent performance (~60FPS at 1080p, most settings on high, no detailed grass, particle settings on low, MSAA+FXAA low) under certain conditions (bumper view and no mirror or map which sucks). The bumper view giving best FPS is similar to NFS Shift 2(on both amd and nvidia). CPU usage at most 50% on any core (tested with VSYNC on). Leads me to believe that optimizations (driver or software) will fix performance.
 
Feb 19, 2009
10,457
10
76

The developers also posted on Steam that AMD drop communications with them well before launch.

They later recanted, Ian Bell had to admit its not true, they've been in constant communications via emails.

Their lead developer also responded initially by telling AMD users to turn down settings. Why? Cos his engineer says PhysX is the problem for AMD.

So you can't pick and choose. Either these devs are incompetent or just plain liars.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126

__________________________________________________________________________
"Project CARS is not a GameWorks product. NVidia are not “sponsors” of the project. "

Oh really? Is this company serious or trolling? :sneaky:

NV 100% disagrees. It is a GameWorks project.

(1) "During Nvidia’s Montreal event – which is underway as we speak – the green team announced that Project CARS will support DX11 (okay, we already knew that), as well as Nvidia’s PhysX particles and Turbulence effects"



(2) http://www.geforce.com/whats-new/tag/nvidia-gameworks

Nvidia GameWorks news feed:
"Gorgeous Next-Gen Racer Project CARS Pulls into PC with High-End Visual Features Including 4K and SLI Support"

__________________________________________________________________________
"We have a good working relationship with nVidia, as we do with AMD, but we have our own render technology which covers everything we need."

Right...

"“NVIDIA has been a great partner of ours throughout the development of Project CARS." said Ian Bell, CEO of Slightly Mad Studios. "We have worked very closely with them on testing, engineering, and driver support and are very excited at the results of the game on PC.”
http://www.geforce.com/whats-new/ar...s-pulls-into-pc-with-high-end-visual-features

__________________________________________________________________________
"we use is a modified version of the same technology we used on the Need for Speed : Shift and Shift Unleashed games, and was entirely developed in-house. The reason the performance drops when there are a lot of particles on screen is simply because processing a large number of particles is very expensive."

Too bad SMS's NFS Shift engine was broken for AMD products from the very beginning. It was always an unoptimized pile of garbage.

Shift 2 was a part of NV's TWIMTBP. Fact is this developer has a history of favouring NV products and spending 0 time optimizing for AMD products.

Let's take a path down the memory lane, shall we!

Developer: Slightly Mad Studios


^ Only the world's most incompetent studio OR a firm that optimized performance ONLY for NV products could manage to create a DX9 game where GTX275 is faster than HD6950 and a GTX580 is 60% faster than an HD6970.

We are talking about the same developer that released a game patch that improved performance of AMD cards by a whopping 45%, titled "Improved ATI graphics card support”." It took them 3 months to do this after the game launched, and AMD literally could do nothing on its end.

This is the developer's patch. How do you release a game in 2011 and release a "game performance patch" that improves performance for 1 vendor by 45% 3 months after launch? That's 100% admittance that the game was not finished/not optimized for all hardware before release. Total incompetence!



^ "The patch really helped out performance wise for me, I think it's good that EA fixed this but it should have never been like this in the first place Nvidia cards ran this game good from the release but I had to wait almost 3 months to play this because the fps were so horrible before. I'll pirate EA games from now on if that's how they want to release games."

__________________________________________________________________________

In case anyone is paying attention, NV's GameWorks will also infiltrate Star Citizen by all accounts.

 
Last edited:

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Also, while everyone in this thread is too busy discussing AMD vs. NV, the large elephant in the room isn't addressed = 960 is just 6% slower than a 780Ti in Project CARS.

Project CARS 1080P 4xMSAA
GTX980 = 52.7 fps
GTX780Ti = 35.8 fps
GTX960 = 33.9 fps
http://www.computerbase.de/2015-05/project-cars-guide-grafikkarte-prozessor-vergleich/2/

In every single GW title, Kepler bombs and it's the same story in Project CARS. $180 960 is just 6% within last gen's NV flagship that cost $700.

This is yet another game, along with The Witcher 3, where GameWorks has ruined performance on Kepler cards.
http://forums.anandtech.com/showthread.php?t=2431768&page=11
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
I thought it was odd but then I looked at Techpowerup's over-all findings with a GTX 660 a 199 MSRP sku up against Fermi's Flagship -- GTX 580. Kepler was much more efficient than Fermi --- the same point for Maxwell -- combined with more driver and developer focus on Maxwell's efficiency advantages or strengths, as the architecture matures.

nVidia's 7900 series didn't age gracefully when compared to the x1900 series from ATI and nVidia's 8800 series. It has happened before.

I don't think there is a conspiracy theory , AMD latest architectural advances have matured more gracefully than Kepler; and nVidia's Maxwell architectural advances are really starting to separate.

If nVidia's focus was trying to sell only Maxwell, they wouldn't off offered DSR support for Kepler and Fermi.
 
Feb 19, 2009
10,457
10
76
SirPauly

I understand the point of not aging as gracefully, look, it happens... but here, this only happens in games that NV is actively involved in during development and with their GameWorks program.

It does NOT happen in neutral or even AMD GE games! Look at Alien Isolation, BF4/Hardline, Civ 5 BE (which does use compute), Star Citizen (780Ti is beating R290X!), Dirt Rally (alpha and excellent performance!).

The question is why is it only happening in games that NV are involved with?

Seriously, if NV is actually involved in the core development of a game, isn't it EXPECTED that the game runs well on all of its current GPUs? They have access to the alpha builds, they can optimize for all their GPUs.

The argument of not aging well would apply if it was the other way around, ie. Kepler tanking in neutral/AMD games, but still running well in NV sponsored games. As that's a situation where non-optimizations in neutral/AMD games can expose its weak architecture, but with NV's active involvement during development of sponsored titles, they can optimize for it.

Think about it. Its the reverse of logic and it implies active involvement of NV during game development leads to a conscious decision NOT to optimize the game for Kepler. In fact, it doesn't imply, because it happens so often already its plain fact.

Remember the 970/980 launch, some of the criticism was that the performance gap was "Meh".. words such as "side-grades" were used. Sure it brought efficiency, but it won't make someone already on 780/Ti want to upgrade for "Meh" results. 50% performance gap is no longer "Meh".
 
Last edited:

destrekor

Lifer
Nov 18, 2005
28,799
359
126
I thought it was odd but then I looked at Techpowerup's over-all findings with a GTX 660 a 199 MSRP sku up against Fermi's Flagship -- GTX 580. Kepler was much more efficient than Fermi --- the same point for Maxwell -- combined with more driver and developer focus on Maxwell's efficiency advantages or strengths, as the architecture matures.

nVidia's 7900 series didn't age gracefully when compared to the x1900 series from ATI and nVidia's 8800 series. It has happened before.

I don't think there is a conspiracy theory , AMD latest architectural advances have matured more gracefully than Kepler; and nVidia's Maxwell architectural advances are really starting to separate.

If nVidia's focus was trying to sell only Maxwell, they wouldn't off offered DSR support for Kepler and Fermi.

Another thing is that Nvidia is often re-working their architectures rather drastically, whereas AMD has found GCN to only need minor reworks to improve with new generations, so their older releases remains viable with current driver optimizations. Which is rather impressive, even if they can't necessarily out-compete Nvidia's performance gains, because they have reached for significant gains of their own through better optimization of their GCN architecture as opposed to needed more drastic architectural changes.

I do look forward to what I suspect will be a new architecture at 14nm, but they could prove me wrong and find ways to make significant leaps with the base GCN architectural approach at 14nm.
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
SirPauly

I understand the point of not aging as gracefully, look, it happens... but here, this only happens in games that NV is actively involved in during development and with their GameWorks program.

It does NOT happen in neutral or even AMD GE games! Look at Alien Isolation, BF4/Hardline, Civ 5 BE (which does use compute), Star Citizen (780Ti is beating R290X!), Dirt Rally (alpha and excellent performance!).

The question is why is it only happening in games that NV are involved with?

Seriously, if NV is actually involved in the core development of a game, isn't it EXPECTED that the game runs well on all of its current GPUs? They have access to the alpha builds, they can optimize for all their GPUs.

The argument of not aging well would apply if it was the other way around, ie. Kepler tanking in neutral/AMD games, but still running well in NV sponsored games. As that's a situation where non-optimizations in neutral/AMD games can expose its weak architecture, but with NV's active involvement during development of sponsored titles, they can optimize for it.

Think about it. Its the reverse of logic and it implies active involvement of NV during game development leads to a conscious decision NOT to optimize the game for Kepler. In fact, it doesn't imply, because it happens so often already its plain fact.

Remember the 970/980 launch, some of the criticism was that the performance gap was "Meh".. words such as "side-grades" were used. Sure it brought efficiency, but it won't make someone already on 780/Ti want to upgrade for "Meh" results. 50% performance gap is no longer "Meh".

There was a shift in the summer of last year with newer titles where AMD was gaining ground on Kepler, especially when Techpowerup changed some of their testing titles. Without Maxwell, Nvidia would be in trouble, imho.
 
Last edited:

ocre

Golden Member
Dec 26, 2008
1,594
7
81
SirPauly

I understand the point of not aging as gracefully, look, it happens... but here, this only happens in games that NV is actively involved in during development and with their GameWorks program.

It does NOT happen in neutral or even AMD GE games! Look at Alien Isolation, BF4/Hardline, Civ 5 BE (which does use compute), Star Citizen (780Ti is beating R290X!), Dirt Rally (alpha and excellent performance!).

The question is why is it only happening in games that NV are involved with?

Seriously, if NV is actually involved in the core development of a game, isn't it EXPECTED that the game runs well on all of its current GPUs? They have access to the alpha builds, they can optimize for all their GPUs.

The argument of not aging well would apply if it was the other way around, ie. Kepler tanking in neutral/AMD games, but still running well in NV sponsored games. As that's a situation where non-optimizations in neutral/AMD games can expose its weak architecture, but with NV's active involvement during development of sponsored titles, they can optimize for it.

Think about it. Its the reverse of logic and it implies active involvement of NV during game development leads to a conscious decision NOT to optimize the game for Kepler. In fact, it doesn't imply, because it happens so often already its plain fact.

Remember the 970/980 launch, some of the criticism was that the performance gap was "Meh".. words such as "side-grades" were used. Sure it brought efficiency, but it won't make someone already on 780/Ti want to upgrade for "Meh" results. 50% performance gap is no longer "Meh".

How many of these games were built completely on new engines built and catered to the GCN on the new consoles?
Games built up from older engines or with slightly updated engines, should fair better.

The industry is moving towards dx11, GCN is the foundation for all the next gen consoles. With keplers 192 cores that all share the same cache, scheduling is very different than GCN. Maxwell layout is better suited.

We can say that Kepler was not the best route or wasnt forward thinking. It doesn't matter, the truth is maxwell SM breaks down into 2 blocks of 64cores that each share cache. Kepler is radically different when it comes to scheduling and balancing the load.

The more modern the game engine, as developers aimed towards maximizing performance on GCN hardware in the console, the more Kepler stands out. Having GCN in the next Gen consoles was always supposed to give AMD an advantage in PC. Is it so strange now that it does? Maxwell is not so radically different from GCN, so it is just better suited.
 

psolord

Platinum Member
Sep 16, 2009
2,092
1,234
136
I installed my GTX 570 as a physx card on my GTX 970 system, and I saw no performance improvement.

Actually the 570 was showing 0% Gpu load, even though it was set as the physx card.

So there's that.
 

LTC8K6

Lifer
Mar 10, 2004
28,520
1,575
126
New 15.5 Driver Coming For Witcher 3 And Project Cars

http://wccftech.com/amd-announces-w...driver-coming-boost-tessellation-performance/

AMD is committed to improving performance for the recently-released Project CARS and The Witcher 3: Wild Hunt. To that end, we are creating AMD Catalyst™ 15.5 Beta to optimize performance for these titles, and we will continue to work closely with their developers to improve quality and performance. We will release AMD Catalyst™ 15.5 Beta on our website as soon as it is available.
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
There was an official response from nVidia, that I missed:

nVidia said:
the assumptions I'm seeing here are so inaccurate, I feel they merit a direct response from us.
I can definitively state that PhysX within Project Cars does not offload any computation to the GPU on any platform, including NVIDIA. I'm not sure how the OP came to the conclusion that it does, but this has never been claimed by the developer or us; nor is there any technical proof offered in this thread that shows this is the case.
I'm hearing a lot of calls for NVIDIA to free up our source for PhysX. It just so happens that we provide PhysX in source code form freely on GitHub (https://developer.nvidia.com/physx-source-github), so everyone is welcome to go inspect the code for themselves, and optimize or modify for their games any way they see fit.
Rev Lebaredian
Senior Director, GameWorks
NVIDIA

https://www.reddit.com/r/pcgaming/c...eworks_project_cars_and_why_we_should/crc3ro1


It's neat that they did free up source code for PhysX, which may improve adoption -- if they hear a lot of calls to free up more GameWorks' middlewares, eventually they may open them up, too, and may improve adoption.
 

tential

Diamond Member
May 13, 2008
7,348
642
121
Wow. I'm impressed with nvidia on that note. I hope they do the same with game works too.
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
I clearly understand why nVidia doesn't desire to give away their hard work and investments but from a consumer and gamer point-of-view these methods or strategies may slow down adoption and desire to see these abilities much more. Would really like to see GameWorks more open, well that is my idealism for more adoption.
 

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
I clearly understand why nVidia doesn't desire to give away their hard work and investments but from a consumer and gamer point-of-view these methods or strategies may slow down adoption and desire to see these abilities much more. Would really like to see GameWorks more open, well that is my idealism for more adoption.


Certainly the idea of game works is good but the anti consumer implementation isnt.

Why can't we just have nice gpu physics in every game for everyone to enjoy, because they want to artificially limit it, then no one can enjoy it...sigh,
 
Feb 19, 2009
10,457
10
76
There was an official response from nVidia, that I missed:



https://www.reddit.com/r/pcgaming/c...eworks_project_cars_and_why_we_should/crc3ro1


It's neat that they did free up source code for PhysX, which may improve adoption -- if they hear a lot of calls to free up more GameWorks' middlewares, eventually they may open them up, too, and may improve adoption.

CPU PhysX only, since March.

You still have to sign an EULA during dev account creation to access it:

"You shall not (nor allow, authorize or assist others to): decompile, reverse engineer, disassemble, modify, sublicense, network, rent, lease, loan, timeshare, sell, transmit, distribute, disclose, publicly display, publish, reproduce, create derivative works based on, assign or transfer the Materials to any other person or entity."

https://developer.nvidia.com/programs/gamedev/register

What does that mean? You can look at the source code, use it it, but cannot modify it. Therefore you cannot optimize it, you have to rely on NV to release updated versions.
 

96Firebird

Diamond Member
Nov 8, 2010
5,714
316
126
CPU PhysX only, since March.

You still have to sign an EULA during dev account creation to access it:

"You shall not (nor allow, authorize or assist others to): decompile, reverse engineer, disassemble, modify, sublicense, network, rent, lease, loan, timeshare, sell, transmit, distribute, disclose, publicly display, publish, reproduce, create derivative works based on, assign or transfer the Materials to any other person or entity."

https://developer.nvidia.com/programs/gamedev/register

What does that mean? You can look at the source code, use it it, but cannot modify it. Therefore you cannot optimize it, you have to rely on NV to release updated versions.

As I have gone over multiple times on this forum, that is a standard software EULA. It does not mean you can't optimize for it, whoever claimed that to begin with has no idea what they're talking about. The code is there for you to see and work with. Of course you cannot modify it, the code does not belong to the user.

AMD's Radeon SDK (which includes TressFX) includes the same wording:

"b. Except as expressly licensed herein, You do not have the right to (i) distribute, rent, lease, sell, sublicense, assign, or otherwise transfer the Materials, in whole or in part, to third parties for commercial or for non-commercial use; or (ii) modify, disassemble, reverse engineer, or decompile the Software, or otherwise reduce any part of the Software to any human readable form."

http://developer.amd.com/amd-license-agreement-sample-code-w_distribution-rights/
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
Rev Lebaredian offered, "so everyone is welcome to go inspect the code for themselves, and optimize or modify for their games any way they see fit."
 
Feb 19, 2009
10,457
10
76
As said, because code was already optimized to start with. NV just needed updated drivers.

HairWorks in Witcher3 tanks performance hard even on NV's GPUs, especially Kepler. One would have to conclude its poorly optimized and awaits NV's updated versions before NV GPU owners can enjoy it.

But besides that, it is a good thing NV is finally making CPU PhysX open source with no obfuscation, it allows all devs to see what's going on in great detail to enable them to better optimize their games. Its progress.

Hopefully in the not too distant future, we'll have GameWorks going the route of open source too so we can have more games running well on all hardware.
 

Hitman928

Diamond Member
Apr 15, 2012
6,123
10,532
136
As I have gone over multiple times on this forum, that is a standard software EULA. It does not mean you can't optimize for it, whoever claimed that to begin with has no idea what they're talking about. The code is there for you to see and work with. Of course you cannot modify it, the code does not belong to the user.

AMD's Radeon SDK (which includes TressFX) includes the same wording:

"b. Except as expressly licensed herein, You do not have the right to (i) distribute, rent, lease, sell, sublicense, assign, or otherwise transfer the Materials, in whole or in part, to third parties for commercial or for non-commercial use; or (ii) modify, disassemble, reverse engineer, or decompile the Software, or otherwise reduce any part of the Software to any human readable form."

http://developer.amd.com/amd-license-agreement-sample-code-w_distribution-rights/

I do notice one explicit difference, in AMD's excerpt it leads to the conclusion that the code can be modified:

AMD said:
Except as expressly licensed herein, You do not have the right to. . .

Do you know what modifications the AMD license allows for? I tried to follow the link but it was bad or the page was moved.
 

96Firebird

Diamond Member
Nov 8, 2010
5,714
316
126
Looks like the link needs to be followed from another link... Go here and scroll to the bottom, click any of the links in the "Downloads" section (I clicked the TressFX one), and the license agreement should be the next page.

After reading it some more, this is just above in section a...

a. Subject to the terms and conditions of this Agreement, AMD grants You the following non-exclusive, non-transferable, royalty-free, limited copyright license to (i) download, copy, use, modify, and create derivative works of the source code version of the Software and materials associated with this Agreement, including without limitation printed documentation, (collectively, “Materials&#8221 for internal evaluation only with AMD processors or graphics products; and (ii) make and distribute copies of the Materials and derivative works thereof created by You in binary code form only for use only with Your products that support AMD processors and in computer systems including AMD processors or graphics products, provided that You agree to include all copyright legends and other legal notices that may appear in the Software.

Does anyone know what this means?
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |