Refine search for CPUs and Memory
Intel Core i7 Performance Analysis
Yesterday Hugo gave us an in-depth analysis of what lies at the heart of Intel's new Core i7 (Nehalem) CPUs, which will be hitting high street shelves in just a few short weeks. Of course, that was all very interesting but no-one's going to buy a CPU based on just the science. All that theory has got to hold up to real-world testing so, without further ado, let's have a look at the numbers.
OK, I lied. There is a little bit more 'ado'. Before we jump straight into the graphs, I would like to set out what our goals were for testing. For a start, it's worth pointing out that, while these Core i7 CPUs will come to be used for all manner of different tasks, they are not server chips so we won't be looking at performance in server oriented tasks. Likewise, we're a consumer oriented site so we feel looking at things like theoretical memory bandwidth and raw integer calculation will be somewhat meaningless for most of you.

Instead, we've put together a selection of 'real-world' applications that we feel are reflective of the sorts of tasks a desktop CPU would be asked to do. Admittedly there are some 'canned' benchmarks in the form of the 3D rendering tests, Cinebench and POV-Ray. However, these are industry standard benchmarks that we feel people value, as they can easily be downloaded and run by anyone.
For comparison we've used Intel's previous top-dog CPU, the QX9770. This still retails for over £1,000 and is unquestionably the fastest desktop CPU currently available. So if the new i7s can beat that chip, they can beat anything. We've also thrown in AMD's Phenom 9950 Black Edition. Although this won't offer much competition to Intel's latest in terms of performance, considering its price it's worth taking a look at how it shapes up.
Our final testing consideration was that of the i7s integrated triple channel memory controller. It has long been known that running your memory in dual-channel mode - whether an AMD system or Intel system - has a performance benefit. So, with i7 using a triple-channel controller, we also wanted to test whether we could see a real-world performance difference by running our system with three sticks of memory. Now, obviously, this isn't an absolutely controlled test as the simple act of increasing our test bed's memory quota from 2GB to 3GB could have some performance impact, regardless of whether that 3GB is configured in three-channel mode or not. Nevertheless, we've thrown in the numbers for a quick and dirty test.

Test Setup
Common System Components
AMD ATI Radeon 4870 X2 Graphics Card
Western Digital Raptor X 150GB Hard Drive
Pioneer BDC-S02BK Blu-ray Drive
Core i7 Test System
Intel DX58SO SmackOver Motherboard
2/3 x 1GB Qimonda IMSH1GU03A1F1C-10F PC3-8500 DDR3 RAM
Core 2 test Sytem
Asus P5E3 Motherboard
2 x 1GB Qimonda IMSH1GU03A1F1C-10F PC3-8500 DDR3 RAM
AMD Phenom Test System
MSI K9A1 Platinum Motherboard
1 x 1GB Corsair Dominator CM2X1024-8500C5D DDR2 RAM
Latest 4 of 12 Comments
Have your say: Leave a comment below about this article.
ilovethemonkeyhead said on 4th November 2008
eyepopper said on 4th November 2008
@Ed, judging by some of the follow-up remarks, I seem to have got my point across.
basicasic said on 5th November 2008
As a once keen overclocker (my first effort was a P75 oc to 90Mhz!!! wow) I'm well past the faster is better mindset. My Q6600 will oc easily to 3.6Ghz with a little extra jui... more
Jay said on 7th November 2008
with multicore processers if you run say 2 single core processes does it use 2 of the cores or just one of them?
See all 12 comments on this article.
Add your comment
You must be logged in to comment. Login or register here.


12 comments
Email this to a friend
TrustedReviews Newsletters
i can see hyper threading being used properly. it certainly improved my old pentium 4. i remember trying my pc (loaded with rubbish, mind) without hyper transport... painful...