Does anyone really care about supercooled LN2 overclocking (other than for the geek sheik factor)? I have an i7 920 system under my desk right now. It's OC'ed to 3.7 on air at just 1.38 vcore. I paid ~$160 for the CPU (nice discount, eh?). It OC's to 4.1 stable if I push the vcore and stress the memory more than I want - so I backed it off a bit to a bus freq of 185 MHz which gives me a nice 3.7 GHz CPU. This is with the stock HSF.
I just ran some Super Pi and it looks like I'm getting 1M runs at around 16 and change (I didn't kill off background processes and I think that's why I'm not hitting as good as some others). But, like others, I don't care much about Super Pi either. My bench of choice is Crysis - and with (only) a single 9800 GTX I'm averaging 38 fps at 1600x1200 with full eye candy ;-)
Personally, I'm not waiting for AMD to come out with more competitive CPUs anymore. I do think I'll grab a couple of their GPUs after Christmas, though as they are just dirt cheap.
Phenom II in a couple of months - no thanks, I've already got my next system (it'll last me for at least 6-8 months). In that time, my next upgrade is to add another (or 3) SSD to play with raid on those (get an SSD - load times just disappear, but that's a comment for a different article)
I just thought I should post this comment because I'm seeing some bad info about the i7 not overclocking. The i7 overclocks easier/better than almost any CPU I've ever seen (and that's a lot). Personally, I care about what performance I can actually get without exotic sub-zero temps on LN2 or phase change cooling. And, I care a lot more about real-world performance that I can get today on hardware that I can buy today. I've already heard whispers of numbers for 32nm Westmere and early ES for SB. The feeling where I work (an MP ecosystem provider) is that AMD is now relegated to the bargain bin and Phenom II isn't going to change that. AMD's manufacturing challenges coupled with a waning architecture and bad management decisions plus the dramatic slow-down in demand for AMD product all spells near and long-term trouble. I just couldn't in good conscious recommend jumping into that mess to anyone. But don't take my word for it - there's plenty of publicly available info about AMD's future tech and future as a business out there.
The only demand I'm seeing for AMD products is for GPUs and (shrinking) demand for Opterons. It's sad, but that's just the state of things right now. Maybe they've still got something up their sleeve to stave off Intel - but it sure isn't going to be this first gen of Deneb/Phenom II. Trust me when I say this - AMD can't pull margins and they are hemorrhaging market share as I write this. I'm seeing virtually no interest in prep build-out for Phenom II/Dragon as all focus is on the volume that Intel commands. What that means is that even if AMD can ramp production of Deneb faster than any launch in their history (and that's not happening), there just isn't ecosystem support. Nobody is willing to gamble any money right now and none of our partners are putting resources into Deneb, let alone optimizing Dragon. It's just too much of a risk while the blue team is offering a much safer bet. That's just the sad truth - at least for the moment.
The Phenom II isn't even going to be readily available for quite a while (word is that CES is just for media launch and it won't hit the channel even in limited qty until late Jan. or early Feb. - and prices will be high compared to expected price drops in core i7). You can get a cheap i7 today (I paid more for my mobo than I did for my CPU) and game away for weeks before Phenom II even shows up at retail.
It's great that we can supercool early ES and get high OCs - but as others have noted, there's little practicality in that.
Just my $0.02 worth - your mileage may vary ;-)
Facebook
Twitter
Instagram