Found this:

http://www.intel.com/Consumer/Learn/family-desktop.htm#/Desktop-des...

Man, my CPU really limits my game play *rolls eyes*

A closer look at the Core i7

Intel Core i7 920 Nehalem 299$
Intel Core i7 920 Nehalem 569$
Intel Core i7 Extreme Edition 965 Nehalem 1029$ (....)

Intel Core 2 Duo E8200 Wolfdale 159$
DDR2 would be fine
prices from newwegg

You know... sometimes I forget just how truly awesome the E8600 is.

I dunno about any one else, but i just dont feel that same impressed feeling that i did with the Core2 generation on their first debut. Im not knocking that these i7s are the fastest to date (cause they are), but the Core2s were leaps and bounds the generation before them. This just felt like a minor tweak to just to stay ahead.

Too many of the desktop application benchmarks do not seem to make use of multiple cores and the numbers do not look right (to me) when I know most CPUs only run at 25%-50% usage with some of their features not taking any effect.

Could TR, please, include 7-Zip benchmark runs into future reviews? I would not even mind of you left out the WinZip benchmark. 7-Zip uses multiple cores, it comes with a built-in benchmark, compresses better than WinZip (at least until WinZip 12.0), handles many file formats, runs on Windows, Linux, BSD and MacOS, and it is free.

( For those who do not know 7-Zip yet: http://www.7-zip.org )

For 3d rendering I have got my 940 running within 15% almost of 2 x E5450.

Thanks for the update. IMO simulations shouldn't be done in the future since the performance can vary quiet a bit in some benchmarks and consumers base their purchase decisions on these numbers.

Thanks for this write-up and benchmarks. I knew there were some differences between the internals of the 965 and the 940, but I didn't expect the differences to be this big. This definitely makes the 920 the most attractive option IMO. It's hard to justify the extra cost of the 940, with such a small performance difference.

Just a little note/correction. The idle power consumption chart has the E8600 using a green bar (AMDs color). It doesn't affect the article at all obviously, just wanted to point that out.

The 920 is a MUCH better deal...

A slightly overclocked 920 will be able to match/exceed the 940 easily, and save you ~$300.

Wow, now I don't feel so hard-pressed with my o/c e8400. Not that I should anyway because of the pricing of i7 right now anyway, but I feel that I can hold on to this chip for a wee bit longer in regards to my predominant gaming usages.

I find it interesting that the main memory latency ends up being right in with AMD's K10 based stuff on the 2.13ghz "uncore" i7 models. I guess AMD didn't do so bad after all. Actually latency tracks uncore clock speed pretty well, right up the chart.

Why doesn't AMD take the uncore to full core clock on BE models? Can their design not handle it? It sure looks like it would pay off. Sometimes 5%, sometimes even 10%... they only wish they could increase the core clock by that much!

Thank you for doing this. It's good to know just how much the difference in the uncore speeds as well as the difference between extreme and performance segment silicon affects performance and power consumption.

Thanks for clearing that up Scott.

The article states on the first page the clock speed of the i7 940 is 2.66ghz. This should read 2.93ghz.

Also I maybe slightly confused here but to simulate a i7 940 wouldn't it have been easier (and possibly more accurate) to overclock the i7 920 instead of downclocking the i7 965? As far as I can tell the only difference between the 920 and the 940 is the core clock speed.

Ooooh....pretty box :)

I just assembled a Ci7 940 system, so I was particularly pleased to see this excellent supplementary report. However, what really stands out is not how much the simulated 940 differed from the real thing, but how much the real thing doesn't differ from the 965. Aside from easy overclocking, which has dubious merit since its already king of the hill, there's just no way to justify the $400+ price premium for an extra 5 - 10% performance.

Thank you for the updated results! It's very much appreciated!

Sure makes me glad i got the 920 model now... i was seriously considering spending the money for the 940... but these tests definitely showed i did the right thing by sticking with the 920. Now all i need to do is wait for the egg to deliver it... **starts twiddling thumbs**