Intel's six-core processors are fast, but enthusiasts almost always want to push unlocked multipliers harder. Core i7-3960X can easily exceed 4 GHz, but what happens to power efficiency when clock rates go up? Sandy Bridge-E demonstrates weaknesses there.
Intel sells the fastest desktop processors you can buy; this much is known. Although some of the company's offerings are disappointingly neutered by locked ratio multipliers, the ones that aren't regularly turn out new speed records thanks to ambitious overclockers. Topping 4 GHz is no problem, even with six cores and a 15 MB of shared L3 cache pushing the complexity of Intel's chips into the billions of transistors. But what happens to efficiency when such a large piece of silicon is pushed to its limits?
Core i7-3960X: Six cores and 3.3 – 3.9 GHz maximum clock rate
It's a good question. As we showed in Core i7-2600K Overclocked: Speed Meets Efficiency, you can actually get better efficiency from this architecture if you overclock it sensibly. Now, we're gunning to see if those results can be duplicated with Sandy Bridge-E, the configuration with six cores, and soon, when Xeon E5 emerges, eight.
Overclocking: For Sport Or Necessity?
Gone are the days when you'd search high and low for that one processor model able to overclock like a beast at a price that was too good to be true. There are so many models now, and so much feature-level differentiation, that it makes more sense to find an affordable CPU that can do what you need it to, and then push from there. For most of us, there's nothing a Core i5-2500K can't do that a Core i7-2600K can for a much higher price. Of course, it doesn't help that most mainstream hardware leads the software industry. Little of what we run on our desktops requires a 4.5 or 5 GHz version of what we already have running at 3 or 4 GHz.
That isn't stopping AMD and Intel from becoming more overclocking-friendly, though (or perhaps it'd be more accurate to say that they're getting more savvy about using overclocking as a differentiator worth a price premium). AMD boasts unlocked ratios up and down its FX stack, for example. Meanwhile, Intel just announced that it will offer, for a small fee, CPU insurance that covers processor replacement in the event of overclocking damage.
Furthermore, Intel finds itself without a competitor in the high-end segment. AMD is currently selling more value-oriented processors, but its best effort currently competes (in terms of absolute performance) with models in the middle of Intel's mainstream portfolio. It can't compete where more affluent enthusiasts are spending money. In terms of manufacturing, Intel is currently about 18 months ahead, which is why AMD's 32 nm-based CPUs and APUs are relatively new, as Intel readies its 22 nm Ivy Bridge-based line-up.
This competitive advantage gives Intel considerable scalability for product planning and efficiency: processors that operate below their design ceiling naturally use less power, giving us plenty of room to measure the effect of tuning for even more speed.
Finding The Optimal Clock Rate
Every processor has an ideal clock rate (or at least an optimal range) at which the chip provides the best possible performance per watt. If you can find that point for your platform, you're sure to get the best performance for the amount of power used. We're using a Core i7-3960X to come up with the ideal combination of low energy consumption at idle with the highest possible clock rate still able to keep energy consumption within reasonable limits.
- Sandy Bridge-E: Does The E Stand For Efficiency?
- Intel's Core i7-3000 Family: Dominating The High-End
- Overclocking: Procedure, Details, And Log
- Screenshot Or It Didn't Happen
- Test Configuration And Benchmarks
- Benchmark Results: Matlab
- Benchmark Results: Professional Applications
- Benchmark Results: Audio/Video And Compression Programs
- Power Consumption
- Efficiency: Single-Threaded (One Core Active)
- Efficiency: Multi-Threaded (All Cores Active)
- Overall Efficiency: Single- And Multi-Threaded
- Sandy Bridge-E’s Efficiency Suffers Significantly Overclocked

Nice to know Intel doesn't just set the stock clock speed for just performance!
Did someone at Intel tell you that was the reason for a lower Turbo Boost limit, or did you just assume it?
I think we should be careful of this kind of guess at another person's, or company's, reasoning. There could be some other cause for the limit - for example, they will obviously sell it for a lower price, so wouldn't a possible reason be they have looser binning specs to allow for chips that wouldn't make it under more strenuous tests through? (Remember, Intel, or any CPU manufacturer, doesn't warrant the product based on what it can be pushed to, and is generally going to provide it at a clock rate they feel is safe over time to guarantee.)
I'm certainly not saying it is a bad assumption, what you said makes sense to me, but I do think there are enough other reasonable possibilities that I wouldn't have stated it as a fact unless I knew it to be.
I do think articles like this are very important; those of us who overclock, especially when we turn off all the power-saving features in hopes of reaching that max stable a CPU can do, should be aware of how much money we are spending if we keep said OC. It's more than just the high end cooling solution.
The people that bash higher capacity PSUs could also stand to learn a thing or two, here. An overclocked CPU can require a huge amount of peak power over and above what a stock CPU needs (349W measured here). An overclocked Sandy Bridge-E and an overclocked GTX 580 could require a peak power of 650W just considering those 2 components!
A Kill A Watt or similar device is a great way to measure how much you actually spend a month operating your computer. You might be surprised.
Tom's Parallel Universe Hardware.
Hence the "probably." Of course, we don't know for sure, nor would Intel ever admit as such, but it's an educated guess nonetheless. =)
Which makes sense since it was written in German =)
/raises beer stein ...
So are all....SOOO impressed by your dangerously overclocked processor. Thank you so much for making that comment.
In other news...
Maybe you guys should have gone backwards a little to see if underclocking would increase the efficiency by a greater factor than the performance loss?
If they all took a bath in liquid Nitrogen then **decent-minded** casual lusr userland would be well-served with the fastest-possible **default** system performance.
Don't feed the animals, Tommy-me-laddie.....
I would agree though be it a somewhat modified multiplier approach, at least for 'my' environment and dependent on 'how' your applications are threaded using 45x/44x/43x. Currently, I'm playing with x48/x47/x46 and Strap values; If you have an ASUS MOBO -- here's a good OC'ing guide -> (scrub to 16 min) http://www.youtube.com/watch?v=Kx2z07sFM2I Again, it all depends on your: 1. vCore (luck of the CPU draw), 2. Thermal (temps/cooling), 3. Applications used.
In addition, I really recommend using 'BIOS Profiles'; example if during the day I'm not doing anything stressful then I'll use a 'Stock' profile, or 'Gaming' profile, or 'Rendering' profile. Each tailored to the environment, a simple BIOS load and reboot you've got what you need from the SB-E.
'My' selection for a limiting factor is the vCore and in essence the heat, I really don't recommend a vCore >1.45v -- so that's my limit. Every SB-E will offer, luck again, different stability per a designated vCore. I also have seen enough data to know both MOBO and Cooling aide signification enough.
Do motherboards allow overclock profiles in the bios so you don't have to manually input new figures to 'turn on/off' overclocking?
6 cores? no one needs them
over 1000$ ? craziness