Core i5-3570K, -3550, -3550S, And -3570T: Ivy Bridge Efficiency

[Guest]

Using a 1000W PSU for this benchmark is beyond overkill. Maximum power consumption is at 155 W - far below 20% where 80+ specs come into effect.

 

[blazorthon]

[citation][nom]dreadlokz[/nom]I'll wait for the revision on Ivy and that shitty thermal desing, and get the 3570k for my new gaming rig =)[/citation]

There probably won't be one. Besides, all it would take is switching out the paste between the CPU and the IHS with either top quality paste or better yet, the fluxless solder used in the other CPUs up until Ivy Bridge. Why revise the whole chip when the problem has nothing to do with the chip and is purely the material between the chip and the IHS?

 

[gam0reily]

there should be an article solely comparing i5 2500k and 3570k. Those are what people call mainstream these days.

 

[gam0reily]

plus, thermal issues on ivy necessitates such a comparison.

 

[jackt]

This Ivy bridge is a piece of crap !

 

[blazorthon]

[citation][nom]jackt[/nom]This Ivy bridge is a piece of crap ![/citation]

Do you have any proof of that? The becnhmarks here on Tom's disagree with you.

 

[oxford373]

why there is just one sandy bridge CPU

 

[oxford373]

ivy bridge launched month ago why toms hardware so late to write this article

 

[akula2]

I don't get it. One many sites, 3770K received positive reviews so why not here?
I agree from the value perspective, probably 3570K is attractive. It was same with 2500K vs 2600K but in *real* world performance I've seen the performance difference between those two.

Would people go for 2700k? I'm confused a bit!!

 

[g-unit1111]

[citation][nom]Sonny73N[/nom]If I build a new rig today, I'd still go with SB. Simply because Intel had replaced the fluxless solder with some kind of thermal paste inside these 3rd gen chips. No wonder IB has heat issue despite their lower TDP compared to SB.[/citation]

I had a ridiculously hard time deciding between Sandy and Ivy for my rig. I really wanted x79 but I couldn't justify the high cost of such a build. In the end I chose Ivy Bridge for its' high clock speed and overclocking potential.

 

[blazorthon]

[citation][nom]akula2[/nom]I don't get it. One many sites, 3770K received positive reviews so why not here?I agree from the value perspective, probably 3570K is attractive. It was same with 2500K vs 2600K but in *real* world performance I've seen the performance difference between those two. Would people go for 2700k? I'm confused a bit!![/citation]

Tom's had a fairly positive review of the i7-3770K. They were simply upset with how it didn't beat its predecessors in overclocking on regular cooling like all other die shrinks have.

 

[SuperVeloce]

yeah, well... it was customary for intel having best models with fluxless soldering. But I remember the outrage with wolfdale (core 2, shrink of conroe) when even E7xxx models came with paste under IHS, because those models were not exactly pentium & celeron money.

If even top models have thermal paste under ihs, it could be a problem with cores when heat is applied for soldering to stick. If to many are destroyed, it could take another stepping to get back soldering, or in worse case scenario, we wait for Haswell.

 

[blazorthon]

[citation][nom]SuperVeloce[/nom]yeah, well... it was customary for intel having best models with fluxless soldering. But I remember the outrage with wolfdale (core 2, shrink of conroe) when even E7xxx models came with paste under IHS, because those models were not exactly pentium & celeron money.If even top models have thermal paste under ihs, it could be a problem with cores when heat is applied for soldering to stick. If to many are destroyed, it could take another stepping to get back soldering, or in worse case scenario, we wait for Haswell.[/citation]

Even if the solder was problematic, Intel could have at least used high quality paste instead of the crap paste that they're using.

 

[SuperVeloce]

yes they should. But it's not an uncommon problem, just look at mobile market. Even if they use thermal compound rather than those thermal soft pads, it's still almost useless. If you change it with a cheap thermal paste from a local computer store, you could get up to 10*C difference on 100% cpu/gpu usage. No ultra expensive compound needed.

 

[blazorthon]

[citation][nom]SuperVeloce[/nom]yes they should. But it's not an uncommon problem, just look at mobile market. Even if they use thermal compound rather than those thermal soft pads, it's still almost useless. If you change it with a cheap thermal paste from a local computer store, you could get up to 10*C difference on 100% cpu/gpu usage. No ultra expensive compound needed.[/citation]

Why does it need to be very expensive to be good paste? Some of the best paste isn't expensive and even then, it's not like the CPUs would need a lot of paste anyway.

 

[blazorthon]

[citation][nom]lawlawlawlawlawyea[/nom]lets compare ivyfail to ivyfail, then the products will look better!!hahahaha this is so funny, they dont even have sandy or bulldozer on there because they would be better[/citation]

Bulldozer would lose in almost everything (if not everything) and Sandy would be right next to them. This article has little to do with Sandy; this is a comparison of the low power Ivy Bridge CPUs. If you want to compare Ivy To Sandy, then look at the main review, not this side article. The only fail here is your misunderstanding of both processors and grammar/spelling.

 

[sammathew]

The finest thing that i liked the most about this article is the schematic explanation of core i5 motherboard.

 

[nikrusty]

[citation][nom]zyzz[/nom]Good article. I think I will get the 3570K over 2500K.[/citation]
Yeah same here...

 

[youssef 2010]

These low power CPUs only target a niche segment of the market. I don't see a real reason behind the introduction of these chips. The S-series Core2Quads made much more sense

 

[Obie327]

Intel made a great chip regarding the I5 3570. Having one myself for gaming is just wonderful. I was looking at the 3770s version myself when they were on sale about a month ago.(K for me!) I'm running my i5 3570K at 3.8 ghz at 1.035 volts and prime stable.(High temps only 63c) Running the lower wattage versions just confirms my Suspicions that there's a trade off when you reduce core clocks/watts. The speed at which a processor finishes a task does in fact save time and energy.

 

[maqsabre]

i guess you should have used deep fritz instead of fritz
as deep fritz is specially designed to exploit multiple cores

 

[oxford373]

i thought ivy bridge CPUs will be more efficient since they implement 22nm and 3d transistor, but they still the most efficient CPUs in the world.

 

[Guest]

Considering that in the real world, nobody turns off their computer after running one item, the efficiency readings should actually be done by measuring all computers at all times, and taking the data based on the slowest. For example, if they finish something in 1,2,3,4,5 minutes, the actual real world efficiency would be 1+4 idle, 2+3idle, 3+2 idle, 4+1idle, 5+0. This would give an actual comparison based on same user workload, and we will likely find that all the computers fit more or less into the same efficiency. While the current workload efficiency method is fine for extremely long work times (3D rendering of a movie, @folding, etc), it does not reflect the typical use of individuals (and IT departments probably could care less).

 

[azraa]

I found it kind of unfair compairing (in games) the 3770K with an aditional PCIe card, when all the others were tested with their IGPs. Besides you are ignoring absolutely the price/performance ratio in this match.

If you take that into consideration, the A8 APU would probably win or be toe to toe with the 3770k
Little bias there Chris? :/

 

[Sakkura]

I found it kind of unfair compairing (in games) the 3770K with an aditional PCIe card, when all the others were tested with their IGPs. Besides you are ignoring absolutely the price/performance ratio in this match.

If you take that into consideration, the A8 APU would probably win or be toe to toe with the 3770k
Little bias there Chris? :/

He did that to compare IGPs to an entry-level graphics card, not to compare the i7-3770K with the A8-3850.

You'll also notice that he included a 3770K without a graphics card in the benchmarks, and the A8-3850 beats that on the graphics side (but loses on physics).