Benchmark Results: Efficiency Testing

We had single core, then dual core, then quad core, feels like 8 cores should be the next big thing.

^6 cores is just another step in that direction of octocores.

wow...doesn't even use up the local power grid back home....
what r u doing wrong intel?
good to see it can overclock nicely tho

Now that you have that i7 Hexacore, why aren't you doing multi-tasking benchmarks. Like running a game and a anti virus, or a game and a browser, or a game, a browser, while doing a anti-virus scan. Or the highly impractical and improbable test of 2 games at once. Seeing as its a hexicore, it should really "shine" during multi-tasking, so why not show us how good it is at that.

+1 to bobfrys

Would really like to see how good this chip can multi-task.

I was really expecting AMD to hit 1st with a hexa, given their good server record. That said, this is more of a paper launch than anything else right now. Performance-wise, the chip sits right where I expected it to sit, so no surprises here.

Right now, there's no incentive for me to swap my i7 920 running at 4 GHz. Sure, as Chris said, i would have loved something along the lines of the afore-mentioned 920, built at the 32 nm node.

Come octocores - I may consider the upgrade. Until then, my fingers are crossed in the direction of AMD - we really need them to keep up, for the sake of a healthy competition.

game + antivirus = choppy performance, antivirus accesses HDD most of the time, causing hangups

game + browser = highly improbable scenario, when you are gaming the browser will just be in idle, when you're browsing, even if the game suffers, you don't care, your focus is elswhere

game + game = whaaaaaaaaat?

The idea is, there are more components in a system, and the CPU is no longer the most important one if you need performance, you really need to speed up I/O (starting with the HDD/SDD). Complex things use the entire system, not just the CPU.

Adding additional programs at once into the equation would probably do more harm than good for the data. In addition to what sihastru said, you could also run into potential program conflicts - even though they may be running in what appears to be the same way every time to you, the user, they probably aren't on the computing side. For one program, the difference is negligible, but that would grow for every program you run. Also, say the programs run at different speeds on different CPUs, where they both need some critical system resource at the same time - then the more powerful CPU would mistakenly look worse, simply because it was faster in the first place.

It is nice to see that several programs can take advantage of 6 cores, even if it's weak. Hardware these days is phenomenal...but humans can still be pretty lazy writing software.

I'll be waiting for octocores as my 920 is still more than fast enough for me. Keep up the good work engineers at intel.

In the majority of these tests there is little or no difference in speed, and the time it would take for the extra efficiency to pay for the price difference, would exceed the lifetime of the chip by many years. Keep the extra cash and buy a SSD this will provide a much higher performance gain and efficiency savings.

i7 975 vs i7 980?? what is to demonstrate??
i think dual socket 17 920 platform vs i7 980 is more appropriate review

because dual socket i7 920 is more powerful and cheap ...and yes regarding power consumption ...the dual socket i7 920 is more efficient...why? because 2 i7 920 must work in full load 4 years to consume enough power(money) to pass the initial cost of i7 980

Looking forward to the time where you can buy a 3 core amd and unlock the potential of a 6 core
Altho i doubt it will ever happen maby with 5 cores turning into 6 cores, picture a lanbox/HTPC when gaming lan unlock the cores, when idle watching movies or playing music remove some of the cores for heat reduction...

Or have i misunderstood something in my equation?

@bobfrys: I can understand game + AV... but game + browsing ? I know plenty of cores makes for a nice epeen, but do we really have to contrive tests so much to pretend and find them any actual use ?

@bobfrys -- I agree

There are many activities which can happen on a daily basis which would benefit from having multiple cores. An example might be number crunching + productivity, I for one do not walk away from my machine while it is encoding, batch processing photos, transcoding, or rendering. It is important that the system remain usable while it is doing heavy crunching in the background. I would even argue that gaming while doing something else is important. For instance measuring broadband performance while gaming, defraging (a different disk) or whatever. The point is that with additional cores, all of these combined activities should become less visible to the operator. Since I switched my personal systems so SSDs I know it has had a huge impact on anything disk related and I can viably have disk activity running int he background while gaming (without impact). This will become even more important as we move forward as there will be a limit to the natural parallelism of some applications and the goal is to multitask. It would be nice to see a multitasking benchmark.

@archange: A paper launch is when someone launches something but does not deliver immediately. I have two gulftowns and four Xeon 5680s sitting in front of me. I have 100 unit quantities of the Xeons on order for the systems we build and expect delivery this month. Hardly a paper launch. The problem with the current set of benchmarks is that very few of the applications used scale well, I am in fact surprised at how poorly Mainconcept scales given the problem it is tackling. For our own use we see 50% scaling with the additional cores so the fact that we get more power efficiency, lower heat, and more tangible performance at the SAME price as the CPU it is replacing is a huge, tangible benefit.

I know that this article was about the extreme edition but I am even more excited about the westmere low power parts which have even lower power consumption and greater turbo capabilities.

Another article I would like to see is the impact of thermal management with respect to the new turbo capabilities of the chips.

@bobfrys -- I agree

There are many activities which can happen on a daily basis which would benefit from having multiple cores. An example might be number crunching + productivity, I for one do not walk away from my machine while it is encoding, batch processing photos, transcoding, or rendering. It is important that the system remain usable while it is doing heavy crunching in the background. I would even argue that gaming while doing something else is important. For instance measuring broadband performance while gaming, defraging (a different disk) or whatever. The point is that with additional cores, all of these combined activities should become less visible to the operator. Since I switched my personal systems so SSDs I know it has had a huge impact on anything disk related and I can viably have disk activity running int he background while gaming (without impact). This will become even more important as we move forward as there will be a limit to the natural parallelism of some applications and the goal is to multitask. It would be nice to see a multitasking benchmark.

@archange: A paper launch is when someone launches something but does not deliver immediately. I have two gulftowns and four Xeon 5680s sitting in front of me. I have 100 unit quantities of the Xeons on order for the systems we build and expect delivery this month. Hardly a paper launch. The problem with the current set of benchmarks is that very few of the applications used scale well, I am in fact surprised at how poorly Mainconcept scales given the problem it is tackling. For our own use we see 50% scaling with the additional cores so the fact that we get more power efficiency, lower heat, and more tangible performance at the SAME price as the CPU it is replacing is a huge, tangible benefit.

I know that this article was about the extreme edition but I am even more excited about the westmere low power parts which have even lower power consumption and greater turbo capabilities.

Another article I would like to see is the impact of thermal management with respect to the new turbo capabilities of the chips.

to be completely honest people that buy this processor, or the 975 more than likely are not going to care about power efficiency, or consumption.

though i have to say, it is interesting to know.

@tomtompiper

if you're talking about why buying a 980x over a 975, it has more cores and its priced at the same $999 probably?