[H] Core 2 Gaming Benchmarks

I'm personally extremely curious about the every day parts overclocking ability. If I can pick up a E6400 or E6300 and then grab another 200Mhz~400Mhz out of it easily then I'll most likely make a big jump from my A64 system to a Core 2 system, though that's one heck of a price for me (I'm still on AGP and DDR, sadly).

I'm extremely curious about the lower end part performance.
 
Skrying said:
Because testing a specific point for the CPU doesnt really paint you a true picture. What's the point of spending $500+ when you only notice a difference in 800x600? Its not like you game at that resolution if you're spending that kinda money on hardware. The benchmarks show you that Core 2 alone is not going to improve your FPS.
Well, yeah? Did anyone claim that Conroe would give a 20% FPS boost at 1600x1200 with 4x AA? The game benchmarks he claims are fraudulent are whatever are pretty basic indicators of CPU bottlenecks in games. For example, if you upgrade your video card more often than you upgrade your CPU, you should probably care a lot more about those 800x600 benchmarks than you do about 1600x1200 when deciding what CPU to buy. Fillrate, PS performance, etc. are increasing much faster than CPU performance, even with multicore CPUs (I only said it so nobody responds with, "but what about multicore chips?"), so you'll hit those CPU bottlenecks faster than you'd think (if you're planning on keeping a CPU for 18 months or more).

And really, there's no such thing as painting a "true" picture. Reviews can't present objective reality; they're opinions supported by evidence (architectural analysis, benchmarks, etc.).
 
Comment on the original article:
I'd just like to say that while I totally disagree with HardOCP's conclusions, I think it's important to have at least one article of this type. Gamers should understand that anything above a midrange CPU typically won't get you much gaming performance. It's also important that people notice that the framerate graphs at that high resolution were spot-on for basically every game.

That said, what lower-resolution game benchmarks do tell you about a CPU is how long it will last. After all, as newer games come out, they will demand more and more processing power to run at acceptable framerates. So you're certainly not buying nothing when you purchase a higher-performing CPU: you're buying longevity.

And, of course, there are other applications where CPU power is much more important, such as media encoding.

P.S. I'm still rooting for AMD. Intel's got this round, it seems, but I'm sincerely hoping for a quick turnaround.
 
Anyone with half a brain who games at high resolutions knows that the GPU is much more important than CPU and using GPU bottleneck scenarios to compare CPU performance is a bad idea.

I do understand what Hardocp is trying to bring across which is basically that if you play at high res with eye candy don't start switching over just yet.

With majority of the new LCD's native at 1280x1024, using that res would give you accurate real world tests (which could still be the same). 1600x1200 and above is not common.

Ofcourse Conroe has other advantages. Not all PC's are used strictly for gaming at all times.
 
On preliminary observation, Anand's article looks quite good and in-depth. Conroe pulls ahead of FX by nearly 30% in multi-gpu benchmarks, something that was needed in this thread. :smile:
 
Agreed; Anand seems to know how to do a proper benchmark. We get power draw, we get "productivity" benches, we get "gaming" benches, and we even get overclocking results. And on top of that, we get to see the full Conroe processor range from 1.83ghz to 2.93 ghz.

Funny how, in gaming, the 1.83ghz Conroe is essentially tied with or beating the X2 4800, especially considering the price point. But that E6600 2.4ghz chip; that puppy is the one to keep an eye one! :oops:

Didn't I just get done mentioning 4ghz on air earlier in this thread? ;)
 
Well I figured I'd add an article (sort of an article) to this thread:

Review of the Quad-Core Conroe (aka Kentsfeild), also OCed to 3.2Ghz
http://www.xtremesystems.org/forums/showthread.php?p=1527913
Only SuperPi (1M in 17.207s), PCMark05 (9556), Cinebench (12s) and some SiSoft benches (53015 MIPs, 36673 MFLOPS) though.

Also, a nice shot of someone really pushing the Conroe at 4.599Ghz with a 1M Super PI of 11.125s.
http://www.xtremesystems.org/forums/showpost.php?p=1572907&postcount=19

Then there is simple insanity with LN2 yeilding 5.204Ghz with a 1M Super PI of 9.875s
http://www.xtremesystems.org/forums/showthread.php?t=106772

This thing is fast.
 
Last edited by a moderator:
One thing to bear in mind is that Intel is on 65nm, while AMD is still on 90nm. I am curious to see what AMD will bring to the plate at 65nm...and hopefully soon.
 
Ah, I'd forgotten I'd seen that.

Unfortunately, it looks like the first rung of 65nm products won't be any higher-performing than the 90nm products (similar to Intel). I think AMD is going to have some belt tightening to do next year...good thing they've saved a bunch of cash.

But, at least once they've moved to 65nm they should have an easier time competing on price and power consumption.
 
Oh, wow... the [H] gaming review was bad. Setting up GPU limited test cases and then proclaiming that a new, faster CPU doesn't do shit is misleading. I'd also say that 1600x1200 isn't a widely used gaming resultion. Try 1280x1024. And what about people with SLI rigs? All HardOCP showed is that CPU speed isn't as important when you hit the GPU bottleneck first. Great, like the [E]nthusiasts this article supposedly caters to didn't know that. Kudos to [H] for delivering the most BS and irrelevant conroe gaming review so far, with the most idiotic introduction ever:
You will see a lot of gaming benchmarks today that just simply lie to you. That is right, you will see frames per second numbers that are at best total BS

Yeah, sure.

I plan on getting a E6600 based system as soon as the CPU is available, plus a X1900XT, which I will replace with a second generation DX10 card when they become available. Since a new video card 18 months from now will likely widen the GPU bottleneck significantly, those non-GPU-bottlenecked reviews are a whole hell of a lot more relevant to me that [H]'s BS article.

Also, [H] cleverly avoided to test games that are CPU limited, just to prove a point that doesn't even exist. How about throwing a few flight sims into the mix? Or how about benchmarking a nice Rome: Total War savegame with two huge armies?

I wonder if [H] will benchmark the next GPU generation using exclusvely CPU-limited benchmarks and then come to the conclusion that the new GPUs don't improve the gaming experience.

I'd have prefered to see a review of that Asus P5B Deluxe i965 board they were using...
 
Last edited by a moderator:
Back
Top