Core 2 Quad Q6600 vs. Core 2 Duo E6850

[Thorsson]

Interesting review on Xbitlabs: http://www.xbitlabs.com/articles/cpu...uad-q6600.html

Their conclusion is that if you OC then Quad wins all the way.

[aidanjt]

It depends on what you do, if an app only loads 1 or 2 threads, then the other 2 cores sit idle, so obviously a higher clock is preferable to 2 more cores. I'm bored of these short sighted reviews tbh, they never show a complete picture of what's going on. Generally speaking a Q6600 is the better buy, they're cheap, fast even for stupid apps, and the amount of applications that are becoming 'n' core aware are on the rise.

[Nemz0r]

It depends on what you do, if an app only loads 1 or 2 threads, then the other 2 cores sit idle, so obviously a higher clock is preferable to 2 more cores. I'm bored of these short sighted reviews tbh, they never show a complete picture of what's going on. Generally speaking a Q6600 is the better buy, they're cheap, fast even for stupid apps, and the amount of applications that are becoming 'n' core aware are on the rise.

Did you read the article ? It showed that Q6600 @ 3.6GHz > E6850 @ 3.85GHz in single threaded application.

edit: and wow @ Lost Planet using 4 cores .
Q6600 2.40GHz 81FPS
E6850 3.00GHz 62FPS
Q6600 3.60GHz 118FPS
E6850 3.85GHz 80FPS.

[aidanjt]

I did read the article, and synthetic benchmarks prove nothing, it's physically impossible for a slower clocked chip to beat a faster clock chip in singled threaded performance even if the slower chip had 36 cores in it, nor did I see any *true* uni-threaded applications there.

[Nemz0r]

I did read the article, and synthetic benchmarks prove nothing, it's physically impossible for a slower clocked chip to beat a faster clock chip in singled threaded performance even if the slower chip had 36 cores in it, nor did I see any *true* uni-threaded applications there.

I meant the gaming benchmarks.

"F.E.A.R. is actually a single-thread game. Nevertheless, the overclocked to 3.6GHz Core 2 Quad Q6600 is considerably faster than the dual-core Core 2 Duo E6850 working at 3.85GHz. This time the graphics driver optimizations played a crucial role."
-Copyright (c) 1999-2007 X-bit labs

For the Quad it's 23.4MHz per frame. The Duo 26.2MHz. So the quad at 3436MHz, would, theoretically have same performance in a single threaded game as the E6850 at 3.85GHz. I think this would be even more profound on the 2900XT, as it's CPU overhead for making stellar use of it's superscalar architecture is larger than that of the 8800 series.

[aidanjt]

That is a lie, F.E.A.R. may not be a multi-core optimised game, but it's by no means a single thread program. A thread may only exist in 1 core at a time, if F.E.A.R. ran everything in only 1 thread it would only utilise 1 core, and thus CPU frequency would be the determining factor, not core count. It's been a *long* time since I've seen a game that runs purely in a single thread.

[Nemz0r]

But you have to remember that even if an application is only using one thread, your graphics card driver is making use of other cores (albeit only a little bit), dependant of what game you're running.

edit: Anyway, whatever the truth, lets not argue: as to your first post we on the most part agree, the Q6600 is a better buy .

[aidanjt]

Bloated/buggy drivers using other cores perhaps, drivers should never eat large enough chunks of CPU time to warrant multi-core setups.