[Gamegpu] GTA V PC Benchmark

[Silverforce11]

Computerbase.de is a good site generaly but currently they are officialy sponsorised by both Nvidia and Microsoft so i would take their reviews with a grain of salt, as always when one site is financiarly supported by any brand.

Well, AT itself had an AMD sponsorship awhile ago. I wouldn't imply that would lead to these major sites being biased.

Certainly some sites have shown themselves over many years to be neutral & objective.

Also, reading computerbase.de's benchmark, they got into a lot of details.. this caught my attention:

"Both alternatives are better looking than that of a rock star. AMD Contact Hardening Shadows know it at the end better to please than the percentage-Closer Soft Shadows. Nvidia has currently still struggling with errors (all shadows are missing on the left field), beyond the shadows are very soft. AMD's version is more detailed."

What does that even mean? Thought such an obvious bug would have been noticed.

 

[Abwx]

I didnt say that Computerbase.de is baised, rather than their results need to be compared to other sites to check their accuracy, as said they do generaly very good reviews, as for the quote you re pointing it s about this feature :

http://techreport.com/news/28114/gt...ved-stable-supports-contact-hardening-shadows

Edit : not rendering the scene accurately will boost the FPs since the GPU is less loaded, not what i would call a fair comparison, i guess that it can be very usefull to have non functional drivers in some instances...

 

[96Firebird]

Edit : not rendering the scene accurately will boost the FPs since the GPU is less loaded, not what i would call a fair comparison, i guess that it can be very usefull to have non functional drivers in some instances...

Is any site even doing this? It would be bad practice to benchmark using settings specific to a brand if trying to compare between the two brands.

 

[Silverforce11]

Is any site even doing this? It would be bad practice to benchmark using settings specific to a brand if trying to compare between the two brands.

Computerbase tested AMD & NV shadows on both hardware. NV runs AMD's CHS fine, even better (!) than their shadows.. but AMD crashed running NV shadows. In-game just fails to load.

Guru3d used Rockstar's soft shadows so its neutral.

 

[96Firebird]

I mean when they do the big performance charts (page 2 of ComputerBase)...

 

[desprado]

Computerbase tested AMD & NV shadows on both hardware. NV runs AMD's CHS fine, even better (!) than their shadows.. but AMD crashed running NV shadows. In-game just fails to load.

Guru3d used Rockstar's soft shadows so its neutral.

Guru3D settings are based on high not Ultra.See there image settings benchmark most of settings are high.

Come on Guru3d u are going benchmark GTX 980, R9 290X and GTX Titan X with high settings?

 

[monstercameron]

People are going to hate this but Watch Dogs is the more graphically impressive open world game.

the 2012 demo maybe...

 

[Silverforce11]

Guru3D settings are based on high not Ultra.See there image settings benchmark most of settings are high.

Come on Guru3d u are going benchmark GTX 980, R9 290X and GTX Titan X with high settings?

Looking at the maxed results from computerbase.de, not even Titan X is playable at 1080p.

Guru3d turned off a lot of fancy shadow options and disabled motion blur from the looks of it, performance improved massively.

http://www.pcgameshardware.de/GTA-5...219948/Tests/GTA-5-PC-Test-Benchmark-1156215/

Also mentions that the shadow options is the major performance loss. They did their bench with NV's PCSS though.

 

[desprado]

Looking at the maxed results from computerbase.de, not even Titan X is playable at 1080p.

Guru3d turned off a lot of fancy shadow options and disabled motion blur from the looks of it, performance improved massively.

http://www.pcgameshardware.de/GTA-5...219948/Tests/GTA-5-PC-Test-Benchmark-1156215/

Also mentions that the shadow options is the major performance loss. They did their bench with NV's PCSS though.

only advanced video option is causing problem and i have seen people on Ultra settings at 1080p on a single GTX 970 and getting above 60fps with MSAA 2X and TXAA on.

 

[96Firebird]

They did their bench with NV's PCSS though.

I thought PCSS didn't work on AMD cards?

 

[DarkKnightDude]

I'm not sure if the Vram bar even works properly in game. I've seen some people saying it counts SLI memory and says you need 7 GBs but they're recording and playing just fine at 80-90 FPS without stuttering or issues.

 

[Alatar]

whose review is to be trusted?

Average the results from all reviews is probably the best idea with GTA V.

E: nvm

 

[Pandora's Box]

I'm not sure if the Vram bar even works properly in game. I've seen some people saying it counts SLI memory and says you need 7 GBs but they're recording and playing just fine at 80-90 FPS without stuttering or issues.

tells me I need 3.8GB of VRAM. When playing the game uses 5-6GB according to Afterburner.

 

[desprado]

tells me I need 3.8GB of VRAM. When playing the game uses 5-6GB according to Afterburner.

It is using for cache and ur min fps drops will be less.

 

[Deders]

Anyone who wants to use Triple Buffering with Vsync on a Nvidia card, enable them both in the GTAV profile in the control panel and disable Vsync in game.

 

[Spjut]

PCgameshardware test
http://www.pcgameshardware.de/GTA-5...219948/Tests/GTA-5-PC-Test-Benchmark-1156215/

PCgameshardware updated the test with CPU benchmarks, checking core scaling on an i7 5820k downclocked to 2,5 ghz

 

[Alatar]

Techpowerup: http://www.techpowerup.com/reviews/Rockstar/GTA_V_Performance/3.html

 

[Makaveli]

Guru3D settings are based on high not Ultra.See there image settings benchmark most of settings are high.

Come on Guru3d u are going benchmark GTX 980, R9 290X and GTX Titan X with high settings?

Maybe there were trying to achieve playable performance instead of a slide show.

Which to me is more important when i'm buying a game.

 

[96Firebird]

Looking at TPU's results, the 295x2 was getting CPU-bottlenecked at 900p and 1080p, didn't want to budge over 108FPS.

Higher numbers than others probably because they didn't use any AA.

 

[desprado]

Maybe there were trying to achieve playable performance instead of a slide show.

Which to me is more important when i'm buying a game.

are u serious?

Even my GTX 960 4GB is getting 40 to 50fps on very high to high settings on 1080p and most of the problem lies on Advanced Video option nothing more.

 

[Hitman928]

are u serious?

Even my GTX 960 4GB is getting 40 to 50fps on very high to high settings on 1080p and most of the problem lies on Advanced Video option nothing more.

I don't see what the big deal is, it's another data point for comparison. Some people (like myself) don't buy high end much any more so it's nice to see a review with settings where a mid range card gets solid performance, it lets me know what I can expect from my card and still have smooth game play. If it's not important to you, that's fine, but it's not like the results are invalid or something.

 

[Makaveli]

are u serious?

Even my GTX 960 4GB is getting 40 to 50fps on very high to high settings on 1080p and most of the problem lies on Advanced Video option nothing more.

Yes i'm very serious.

and considering you own a 960 why do you care what numbers the titan X or 290x or 980 are pulling what is the benefit to you?

 

[Jacky60]

Come on ladies 4k is where it's at these days, ideally at 40 inches ladymen! I'm stuck on FXAA but it does look very good and a consistent never stuttering 60fps/60hz with Vertical sych frame pacing etc, all other settings maxed but I clearly need more oomph somewhere. It's on a WD 1GB HD so I suspect with a couple of 850pros in raid it may speed up. CPU and GPUs rarely get above 50% with all cores similarly loaded.

 

[desprado]

Another benchmark at ultra settings
https://www.youtube.com/watch?v=x_wIsvmbMoc&hc_location=ufi

 

[desprado]

Yes i'm very serious.

and considering you own a 960 why do you care what numbers the titan X or 290x or 980 are pulling what is the benefit to you?

Because GTX 960 is just for physX and i game on GTX 980 SLI.