Core i7 vs. Phenom II X2 vs. X4 scaling performance analysis
The review that we are presenting today evaluates the GTX 480, the HD 5870 and HD 5870 CrossFire with three CPUs from 2.6 to 3.6 GHz. It is an update to last November’s “Performance Meets Value” series where we used our GTX 280 and the HD 4870-X2 to benchmark our then brand-new Intel’s Core i7-920 processor against the Phenom II 550 X2 and 720 X3 and also against our Q9550S at speeds up to 4.0 GHz. We concluded that series where we additionally benchmarked the Phenom II 955 X4 with our HD 4870-X2 and HD 4870-X3 Tri-Fire in “Core i7 vs. Penryn vs. Phenom II with HD 4870-X2 & TriFire” which we published this January, 2010.
Here is our conclusion from almost exactly one year ago and today we want to see if anything has changed over the past 12 months:
If you are looking for bang for buck now, with a single video card of the HD 5870 class of card or HD 4870 CrossFire, Core i7 is generally overkill for gaming and often it not the fastest when compared to either our (now midrange) Q9550S or even the decidedly budget Phenom IIs. Of our fifteen gaming benchmarks, even with our HD 4870-X3 Tri-Fire setup, only World in Conflict seems to really benefit from a quad-core so as to make any practical difference to the game play – and Core i7 was also the fastest.
The rest of the benchmarks demonstrate that all five of our CPUs run fairly close in performance to each other and the results are satisfactory once they are overclocked a bit. We even note that overclocking is not necessary if you don’t mind sacrificing just a few frame rates. Our current conclusion from this and past testing would still confirm that graphics are the single most important factor for determining most game frame rates at maximum detail – not the CPU. As long as you have a decent tri- or dual-core, you are not really losing much by not having four CPU cores – other than bragging rights – in most games.
Our current performance versus value series has drawn to a close – for now. We will revisit it in the future when we have much faster graphics than our HD 4870-X3 Tri-Fire setup – perhaps with Fermi SLI or with HD 5870 CrossFire, this coming year.
That future is now, and we are expanding on our old review with more games to benchmark and powerful graphics: HD 5870, GTX 480 and HD 5870 CrossFire. And this time for our three competing CPU platforms, we are going to include AMD’s 955 X4 and 550 X2 to test against our Core i7-920 at 3 core speeds. That means that we are now benchmarking with three CPUs at three clockspeeds each; 2.6 GHz, 3.2 GHz and 3.6 GHz. In over a month of solid benchmarking – literally hundreds of individual benchmark runs which we have gathered, analyzed and charted for you – we now have some more solid evidence which finally answers the question as to what kind of CPU is needed for really fast graphics at the end of 2010.
We are continuing to test at two of the most popular demanding wide-screen resolutions, 1680×1050 and 1920×1200, plus at 2560×1600. We use 4xAA or 8xAA plus 16xAF and with maximum DX11/10.1/10/9c details whenever it is available and with our CPU underclocked to 2.6 GHz, at 3.2 GHz and overclocked each to 3.6 GHz to demonstrate CPU scaling with framerates. We are particularly going to pay attention to our HD 5870 in CrossFire compared to the HD 5870 and GTX 480 by comparing framerates with our 3 CPUs at three clock speeds each. We have also condensed our usual charts so that only one is used for each resolution that we test at.
We are benching our three CPU platforms with the GTX 480 our HD 5870, and HD 5870 CrossFire as the fastest graphics available to us this last Summer when we did the testing. Although this review got delayed by the flood of new hardware that ABT has received for evaluation over the past few months, it is still just as up to date as if it were benched with the latest drivers from this month as it will clearly show relative performance with CPU cores and scaling.
For all of our testing in this review, we are benching with Catalyst 10-6 and GeForce 257.21 WHQL drivers; each one a very solid and stable driver set. We will now use our highest performing Phenom II, 955 X4, in addition to 550 X2 so as to test AMD quad- and dual-core CPUs against our quad-core Core i7 920 with HyperThreading on. So now we have a very high-performing and rather expensive Intel Core i7 system to set alongside the decidedly value AMD Phenom II X2 and X4. We will test each of our three video card configurations with each of our three CPUs at 2.6 GHz, 3.2 GHz and at 3.6 GHz for a clock-to-clock comparison which will demonstrate framerate increase with CPU scaling.
When we say “performance meets value”, we mean that the Core i7 X58 motherboard, CPU, and tri-channel DDR3 is much more expensive than its Phenom II counterpart. We will also note that the AMD motherboard’s PCIe graphics bandwidth is limited to 8X + 8X which is half the PCIe bandwidth of the Intel motherboard. So, we naturally ask, “is it worth it” to buy the i7 system even for really fast graphics over the Phenom II system; and is quad-core necessary for med/high-end PC gaming?
Please continue on to the next page for the complete hardware and software setup of our platforms – AMD’s Phenom II versus Intel’s Core i-7. We shall see what happens when high performance finally meets value in PC gaming with fast graphics at the end of 2010.
Flat-out amazing!!! I’ve never seen anything so epic like this. So, there was not any microstuttering in any of the above games, where the “measured” 40 fps appeared to look more like “perceived” 25 fps? I guess microstuttering is not noticeable if the measured fps is above 70-80, since half of this (45 fps) would still appear to be relatively smooth. Anything below 30 fps becomes really noticeable, so were there ever 40-50 fps instances with 2x 5870 CF that felt like 25fps or so? I’ll take your word for it, if you were actually watching 10,000 hours worth of benchmarking, ha (just kidding, don’t shoot my head off)!
Surprised there are not more comments. For a single video card it does look like a dual core is more than enough. Sure there are a few games that take advantage of four cores, but the fact remains they remain in the minority.
Bobert, I’ve been saying the same thing for months now.
Wow, very thorough and detailed article. It’s one thing to test CPU performance using a single video card, but it must take some brawn to do it for three different video configurations.
There’s so much data here to look at in so many ways. I suppose if you would have included an SLI setup we would then be able to determine how CPU speed affects SLI vs. Crossfire. If I’m looking at this data right, though, it seems Crossfire sees benefits from quad cores more than single video cards do.
Far Cry 2 seems to be a good example of this. Also Far Cry 2 shows interesting relationships between CPU and the GTX 480. The single HD 5870 doesn’t really react to CPU speed and cores the same way dual 5870s and the GTX 480 do. That’s pretty interesting.
So well done. If I only had one suggestion is that I would like to see GTA4 benched, mainly because I own it and good, thorough, and updated benchmarks of it are not easy to come by.:)
Concerning the Far Cry 2 numbers, despite being beaten with faster processors, the HD 5870 paired with the 2.6 dual core is actually outperforming the Crossfire and Nvidia setup. That’s what I find a bit interesting here.
AWESOME REVIEW.
This is EXACTLY what review websites ARE NOT putting out.
A non GPU-bottlenecked review showing how i7 really does have a significant gaming lead over Phenom II.
God you’d be surprised how many AMD fanboys still believe (and spread rumors) that Phenom II is plenty for 5870 crossfire. Psh. Plenty on today’s games maybe, but that is due to the PC gaming community being SNARED by the noob console community and their half a decade old setups.
Ok I’m ranting.
Two Thumbs!!
Raidur means it shows how Phenom II bottlenecks 5870 crossfire.
Everyone knows i7 is faster in games.
PS. I’m not Raidur.
PSS. I’m Raidur.
PSSS. Or am I?
Awesome review GJ.
Please let me know if you’re looking for a article writer for your
site. You have some really good articles and I believe I
would be a good asset. If you ever want to take some of the load off, I’d love to
write some material for your blog in exchange for a link back
to mine. Please shoot me an email if interested.
Thank you!