[joe]

All of the 27" iMac game benchmarks I've found so far list Quad-core R4850 models along side Dual-core models with slower R4670 graphics. There are even CTO Quad-core 2.8GHz benchmarks with R4850 but never a CTO Dual-core with the better graphics chip. Does anyone here own a CTO dual core with the faster graphics? If so, could you post a few gaming benches?

Yes, the quad cores have turbo boost and will increase speed from 2.66GHz (or 2.8GHz) automatically. And I know some games are GPU-bound rather than CPU-bound. I'm just curious how the 3.06-3.33GHz dual-cores compare with as many 3D games as possible when using the same graphics.

[mduell]

The quad is so cheap once you put the 4850 into the dual that it's a no brainer.

[P]

You can probably infer the performance from PC benchmarks about which games are GPU bound and which are CPU bound, but as mduell says, the quad is a no-brainer.

Let's be clear here: The Core i5 will be faster than either of the Core 2 models in any reasonable usage scenario. Yes, you can create freak situations with singlethreaded code running inside the L2 cache where the Core 2 3.33 GHz can beat the Core i5, but those are really weird cases. Real world is heavily in favor of the Core i5, even with all current software optimized for Core 2.

The only remaining reason for Core 2 then is to save money - which is fair enough, not everyone needs the power of a quad - but in this case you'd only save $150, or about 8% of the total price, for a significant decrease in performance. That just doesn't make sense.

[joe]

I was asking about games, not pro apps. Before I started this thread I'd already found several iMac Game benchmarks where the i5 and i7 iMacs were either slower or equal to the previous gen 3.06GHz Core 2 iMac with the same R4850 graphics:
iMac Core i7 - 3D Apps at 1920x1200

That's not to say there aren't exceptions (ex X-Plane). But it's pretty obvious that most games don't take advantage of 2 cores much less 4. And what about turbo boost?! The i7 should've ramped up to 3.4GHz and blown away the older Core 2 iMac rather than being 11fps slower. Could that have been heat related? Does Turbo Boost throttle back under certain conditions even when not using 4 cores?

One way to find answers would be to add a 3.33GHz iMac w/R4850 to the mix.

[P]

The only test where the C2D model was faster than the newer ones was ET:QW. That test also had the i5 and i7 being equal. That's a very suspicious test - most likely something was off there, and the game is just GPU-limited - which the author of that article even notes. In general, most new games are GPU-limited with a Radeon 4850 at the resolutions a big iMac runs at. Most likely the Core 2 was nowhere near loaded, and the i7 was probably more or less idling. It was certainly nowhere near having to use turbo.

If you really want to save $150 and get a computer with less than half the CPU power, then sure, go ahead. I think you'll regret it soon enough, as newer games are better at using more cores, but it's your money. Absolutely do not buy the 3.33 GHz model, however.

[joe]

Quit assuming that business about the $150!. The thread was started to gather actual game benchmarks for a new, high end and (MORE IMPORTANTLY) untested iMac. Regardless of whether I agree with RobArt's conclusions (and I agree), it would still be beneficial to actually test the darn thing.