UPDATED: AMD never did hold the highest 3DMark record
Opinion - Santa Clara (CA) - An Intel employee and well-known overclocker contacted TG Daily to point out that AMD actually never did hold the 3DMark record. FutureMark's website shows the 45474 3DMark AMD machine result was submitted on January 12, 2009. However, results from an earlier Intel Core i7 powered machine came on January 4, 2009 (8 days prior) and show a score of 46644, a full 1090 3Dmarks higher.
TG Daily quickly observed, however, that this is not the main issue at hand. While the Intel machine clearly has the highest 3DMark rating per Futuremark's site data, and the AMD folks were either negligent or unaware of the Core i7's score when they claimed the highest (see our previous coverage), the reality is that wasn't AMD's big triumph.
While their 45474 3DMarks is truly nothing to hold in ill regard, AMD's big triumph was being able to take their normally clocked 3.0 GHz Phenom II X4 CPU up to a whopping 6.5 GHz using liquid helium (not liquid nitrogen), while operating well below -200 degrees Celsius. This 117% overclock was the big news in the story.
There's also some discrepancies here. For example, the AMD entry on Futuremark Intel pointed us to only shows a 4.481 GHz clock speed (49% overclock), and not the 6.5 GHz shown in the video. But even more interesting than that, the Intel Core i7 entry on Futuremark Intel pointed us to shows their Model 940 2.93 GHz CPU only clocked at 3.693 GHz (a 26% overclock). In addition to the lesser percentage overclock, the machine in question used ATI Radeon HD 4870 X2 video cards to generate the scores - an AMD product. In addition, the 46644 score was achieved using drivers unapproved by Futuremark.
Intel has yet to show us anything significant in this overclocking arena with Core i7, other than the fact that it can overclock by 26% and produce a higher 3DMark score than AMD's 49% overclocked machine while relying on an AMD/ATI graphics card - which admittedly does speak to Core i7's greater performance per clock (PPC). But even so, I must now ask you, oh Intel employee, where's the 2.93 GHz Model 940 overclocked by 118% (one percentage point higher than AMD's overclock) at 6.387 GHz? Or better yet, where's the overclock to 6.51 GHz (a 122% overclock) showing your true dominance and superiority in x86 space?
Our readers want to know, Intel: How much game does Core i7 have? I mean, it's fine if you're going to rely on your better PPC and mild overclocking abilities coupled to AMD's graphics cards (which ultimately do the hard work for you anyway) to give you those big scores. And seriously, just so you know, we're really okay with that because we know it might be the best you can do. :)
The opinions expressed in this commentary are solely those of the author.
UPDATED: January 29, 2009 at 5:20pm CST
Several commenters have indicated the clock speeds reported on Futuremark's front-page are not accurate. I'm not sure why this is, but apparently it is a well-known phenomena. The article reports 3.693 GHz and the description reports "I7 940XE BO @5203mhz". As such, at face value I retract the portion where I claim specific percentages in overclocking. According to the comments, the Intel chip actually did clock at 5.2 GHz, and not 3.693 GHz as stated. That would represent a 77.5% overclock, greatly exceeding AMD's reported 49% overclock. However, it is still well below AMD's reported 6.5 GHz 117% overclock.
As for the other commenters:
I'm amazed how many readers have attributed this article to fanboyism. Its intent was just to "talk smack" at Intel, and be in the general framework of humor. The article was posted as an "opinion" piece, and not news, and for that reason. It was supposed to be a somewhat comical way to stir up a little competition between AMD and Intel for the overclocking crown. I hoped Intel might come here, read this article and say "We can't let AMD have the 117% crown," and then go off into a lab somewhere and come back with a 250% overclock. That's all I was going for.
But I can still be big about this: To all of our readers who posted or submitted comments saying that I'm a victim of massive AMD fanboyism ... I apologize. It seems clear that I should've written the article in a different way so the "smack talk" aspect would've come across better. I thought it was clear, bounced it off another TG Daily writer before publishing, however I can see now it went right past some of our readers.
I hope now this has been corrected. And I hope if you re-read it with this thought in mind you'll see what I was going for. I look forward to your comments.