[MIG-31]

Grand Theft Auto IV for the PC is finally here! It has a steep set of recommended system components, so we're going to take it apart and tell you what you can expect from seven of today’s most current and popular video cards.
____________
Source: [H] Enthusiast

[OmegaRED]

Wow! Just the article I was waiting for. I'm only about halfway through it but it's reassuring to see that the 192 and 216 core 260GTX's are indistinguishable in performance and it is possible to enjoy this game at high settings provided you have a beefy quad core 4gig system.

I can't help but chuckle looking at the IQ section, the lack of AA and AF is a killer on image quality. If this game ever gets AA support whether forced or in game it'll look ten times better.

[BlueMak]

At one point they say that it is a modest game in requirements, then say we can hope to aim for 30 frames per second, but 25 is good too...Then they get a newish quad core CPU at 3.6Ghz. Funny. While I understand they sort of mention it is going to be a graphics card test, it sure feels like (title too) that it is GTA IV gameplay performance and IQ test. Shouldn't they use a bit more reasonable system? How many people have the specific CPU? This "test" doesn't help at all, even for a graphics card comparison since I am sure some people will buy the game and graphics card that looks better, then realise their core 2 duo or amd x2 is laughably slow for this game.
Their last page is better, but still, I expected more from [H].

[OmegaRED]

Originally Posted by BlueMak

At one point they say that it is a modest game in requirements, then say we can hope to aim for 30 frames per second, but 25 is good too...Then they get a newish quad core CPU at 3.6Ghz. Funny. While I understand they sort of mention it is going to be a graphics card test, it sure feels like (title too) that it is GTA IV gameplay performance and IQ test. Shouldn't they use a bit more reasonable system? How many people have the specific CPU? This "test" doesn't help at all, even for a graphics card comparison since I am sure some people will buy the game and graphics card that looks better, then realise their core 2 duo or amd x2 is laughably slow for this game.
Their last page is better, but still, I expected more from [H].

HardOCP (and a few other enthusiast sites) always pick the fastest processor available and then try to overclock it to remove the cpu limitation from testing; it limits the appeal of the benchmarks but I think GPU prices and power have caught up enough that most of us can run games at the settings [H] benchmarks with.

As for the 30fps playability that was done for Oblivion and Crysis so I think it's fair to do it with GTAIV especially since that's higher than what the console versions ran at.

I have a feeling that [H]'s testing of GTAIV will need to be updated once SLI and Crossfire support is added. This should create a situation where SLI 8800GT's and 4850's can run at playable settings regardless if it's dual or quad core.

Glad he mentioned the crashes (maybe they should benchmark how often it crashes with each GPU!) because I tried to have a game recently and it kept crashing.

[Alex]

I agree with blumak, its not thorough at all. If you are doing a game analysis of a product like this then you at least need to include a dual core to show the gains from buying a quad core.

Its useless as an overall analysis of the game because unless you already have a quad core and can OC, then the test system is just a basic idea of "what is possible". There has been so much debate on this game and what is needed that why on earth wasn't it a priority to show that?

what about the kid who has a dual core, gets money for christmas and has yet to buy the game? he won't have a clue what any of it means subjectively.

I dont like their gaming articles at all, lack of thought. They get more like tomshardware every time I read them.

[sterben]

I can see why you did quad since its supported but no dual? Come on, not everyone has a quad.

[OmegaRED]

Originally Posted by Alex

what about the kid who has a dual core, gets money for christmas and has yet to buy the game? he won't have a clue what any of it means subjectively.

I dont like their gaming articles at all, lack of thought. They get more like tomshardware every time I read them.

Do we really need more benchmarks telling us what hundreds of people have already told us when they posted their benchmarks on forums: "Dual cores aren't enough!"?

I'd rather see testing done which shows what other hardware I will need to play GTAIV with the video card I already have. Seeing as how [H] included all of the most recent video cards (likely to be a xmas gift for any kid) they did a good job. Duals aren't even the recommended specs for the game.

I guess it would have been nice to see some older hardware thrown in for comparison but that's not the MO for [H], they test with enthusiast settings for the enthusiast crowd not the kid putting a 280GTX into his 3 year old Dell

[OmegaRED]

Here's some benchmarks of GTAIV Cpu scaling

PCGH - Review/Benchmark: GTA 4 (PC): CPU benchmark review with 13 processors

Maybe this is why HardOCP didn't use a different CPU

GTA 4 (PC) - Problems when changing CPUs
Although we uninstalled GTA 4 inclusive Securom, Games for Windows Live (GFWL) and Rockstar Game Social Club (RGSC) and "revoked” our license before we changed the CPU of our system and reinstalled the game afterwards, GTA 4 could be run anymore. Therefore we had to change the hard drive with the operating system on it to get it running again. Unfortunately was GTA 4 refusing to start once more as we changed the clock speed of our Phenom CPU.

Because of this we can't provide any results for Yorkfield and several Phenom CPUs at the moment. We contacted Rockstar and the processor producers in order to find a solution for the problem.

And the older GPU's

As GPU power is concerned, we can give an all-clear. Even at 2,560 x 1,600 (but also at 800 x 600) a Geforce 9800 GT is only slightly slower than a GTX 280 - GTA 4 primarily requires CPU power, reducing the resolution doesn't result in much more fps. On two test systems we were able to get a Radeon HD 4870 running. The benchmark revealed that this card also delivered similar results.

[brutusmaximus]

Originally Posted by OmegaRED

Do we really need more benchmarks telling us what hundreds of people have already told us when they posted their benchmarks on forums: "Dual cores aren't enough!"?

Yes we do. If we went on 10% of the information posted by kids on forums then we would be in a bad place. There is absolutely no logical reason to handle a game analysis and not detail the improvements or differences by investing in a quad core. Steam research shows the largest percentage of userbase is actually dual core, not single core. To assume that running tests solely on quad core is helpful with this particular editorial is foolish. After all why not help sell the higher end products if they really are that beneficial?

I'd rather see testing done which shows what other hardware I will need to play GTAIV with the video card I already have. Seeing as how [H] included all of the most recent video cards (likely to be a xmas gift for any kid) they did a good job. Duals aren't even the recommended specs for the game.

Well I don't assume that a hardware enthusiast site should then be doing PR for Rockstar. If the biggest percentage of userbase is using Dual core (which I would say is accurate even for DH or H readers) then it needs addressed.

I guess it would have been nice to see some older hardware thrown in for comparison but that's not the MO for [H], they test with enthusiast settings for the enthusiast crowd not the kid putting a 280GTX into his 3 year old Dell

Very harsh and a little silly. the Hardware MO for Hardocp seems to be that everyone should have a 4870 or 280 GTX with Quad Core. why not show what a high end dual core will handle with the same cards first then show the differences. its logicial and more thorough, but as it is Hardocp its to be expected. They long ago stopped being thorough and are getting lazy in their old age.

I hope DH do it right, time permitting but Zardon said in PM they dont have time right now.

[OmegaRED]

Bah, I think you guys are being too harsh towards [H]. They test the same way for all games and it's one of those love it or hate it methodologies but if they didn't do it that way who would? We've got DH and so many other sites who will benchmark with a broader spectrum of hardware and [H] fills the role of the uber specs pc benchmarking for those who are into that sort of thing.

[macbiter]

Originally Posted by OmegaRED

Bah, I think you guys are being too harsh towards [H]. They test the same way for all games and it's one of those love it or hate it methodologies but if they didn't do it that way who would? We've got DH and so many other sites who will benchmark with a broader spectrum of hardware and [H] fills the role of the uber specs pc benchmarking for those who are into that sort of thing.

Well it is rather clear you are a fan of Hardocp and thats cool, I quite like their site myself. I will however agree with the other people in this thread to a certain extent to air my views that it isnt really an "analysis" but more a "showcase" of high end hardware. An analysis by definition should clearly depict to the readers the subject matter performing under as many variables as possible. Within the contexts of these parameters, the article is sadly lacking in execution, thoroughness and a passion for the audience.

I think its also a little unfair to be dismissing sites like DriverHeaven as just catering for the masses. As far as I can remember they have lead the way in many areas, with one coming off the top of my head immediately. The first place to get Skulltrail running with 3 280 GTX's when other publications failed to do so (and im not even mentioning the highest end power supplies on a regular basis, the most thorough GFX reviews as well as even Samsung monitors at 4k!). If thats not ultra highend im not sure what is.

hardocp need to refocus and perhaps steer away from publishing video content with kyle showing various nonsense items like "sata cables" (yes really) and concentrate on getting more thorough testing handled.