[Eldonko]

Yah interesting stuff. I wonder how hot next gen will be..

Originally Posted by hokiealumnus

Superb link from eldonko @ XS (thanks man): http://youtu.be/bm6ScvNygUU . Very much well worth watching and applicable to the increased temperature issue.

[MattNo5ss]

If you can't RMA the 2500K, then I'd go 2600K for 24/7 folding.

[Eldonko]

Unless you fold at stock then 3770k will run cool and save a ton of power.

[moocow]

'Dat IMC

[rustyfender]

i was folding at 4.8 on the 2500K and it stuck around 52 deg. from what im seeing folding at 4.8 on a 3770K would put me in the high 80s. eeek. but eldonko has a point it will prolly save quite a bit of power.

hmmmmmmmm

[Eldonko]

I think I would go with a 3770K for folding and a bit of a lower OC. Even at 4600Mhz it should produce what a 2600K does at 4800K (or close) and since you run 24/7 it will make a difference in power bill in the long run.

Good read here on what you can expect for power savings: http://www.hardwarecanucks.com/forum...review-22.html

[rustyfender]

thanks for the aricle. good read!

[diaz]

Anyone UVOC yet? whats the lowest voltage you can be stable for 4-4.2 ghz?

[hokiealumnus]

Undervolting and Overclocking on Ivy Bridge.

[EarthDog]

LOL @ UVOC again... for those that may not know: Under Volt, Over Clock.

[muddocktor]

Nice review hokie! I can't wait to get my retail 3770k when they go on sale to give it a whirl on water and various high end air coolers.

[Janus67]

Originally Posted by hokiealumnus

+1.

Jman13, your point is well taken but isn't applicable to our readers, who live in a world where no one measures processor temperatures in Kelvin. If I had broken out Kelvin in the review, there would have been twenty of your posts instead of just one from the other perspective.

Also, unrelated to temperatures, everyone should go check out Shamino's magic. Remember how I said the IMC on these things is insane. Look what happens when you pair it with equally insane memory on LN2.

[Jman13]

Originally Posted by hokiealumnus

My statement is correct. So is yours. The hottest core measures 46% higher in degrees celsius than Sandy Bridge. That is a correct statement. You may be correct, on a Kelvin scale, the heat increase is only 7.8% in terms of kelvin, but if you're going to be technical, I will to.

But it's not a correct statement. Sure, you can say the number in C is 46% higher...but that's a meaningless statement. Let's say you'd measured in F (a scale most of your readers are very familiar with). Your numbers would have been 180F for IB and 133F for SB. So that means the measurements are 35% higher. Which one is correct?

If you don't want to use Kelvin in your article because you think your readers would look at you with a blank stare, then don't use percentage increase on heat, just say that IB runs 26C hotter. It's not a difference in stating things, it's a flat-out inaccuracy. I mean, let's say we had processors that ran at 1C at load, and a new processor had the EXACT same heat increase as IB (7.8%). That would yield a Celsius temperature of 22.4C...an increase of 21C. Would you then write in your article that the new processor ran over 22 times hotter? That it measured 2,200% higher? How misleading would that be? Just because the wrong numbers are less sensational due to the slightly higher starting temperature doesn't mean that they're any less wrong in this case. In this example the actual increase in temperature is identical to the real IB vs SB numbers.

This is not a discussion of 'choice of units' it's a flat out incorrect (or, due to the way you worded it, a horribly misleading) statement.

Anyway, I'm sorry I'm being the adamant jerk here. The review was very well written overall, and I quite enjoyed the article. It's just the engineer in me that has to speak up when I see something like this.

[robble]

Jman, I appreciate the lesson. It's something I never would have thought about. So the lesson here is to not think about percentages and just go with how many degrees higher.

However, I don't think anyone was trying to mislead us by stating percentages. It also works for us ignant laymen =) We don't need the precision an engineer does. From now on though I'll ignore remarks about percentages on temps and just go by the number of degrees higher.



Now back on topic. IB just isn't exciting me as much as I hoped it would. my i5 750 @ 4Ghz seems to be doing everything I do just fine. MAYBE it's bottle necking my 7970 in some games but everything already runs so smooth I'm not noticing it. I am the kind of person who caters to wants more often than needs though so if there had been a big jump I would have bit. I can see where benchmarkers might get excited but I doubt I'll ever go below ambient.

So now, how long until intel's next chip?

[jd101]

so will a i7 3770k do 4.2 with a h80 with out cooking its self ?

[EarthDog]

Likely, yes.

[Brolloks]

Originally Posted by jd101

so will a i7 3770k do 4.2 with a h80 with out cooking its self ?

It should run fine, closer to 4.5 Ghz you might hit the high temps.

Being a 22 nm I suspect running temps above 65C is not good for longevity, maybe Hokie can guide as there?

[Bobnova]

Hokie is absolutely correct to use C in my opinion, CPU temps are measured in C by everybody, including Intel and AMD the people that make the CPUs.

His usage of percentile is 100% accurate as well, it was already defined as being in C.

If you want to be pedantic about it, you've chosen the wrong thing to be pedantic about. Rather than the temp scale being used, go for the percentage itself. Regardless of the scale used you cannot accurately use any method to estimate your eventual temps without knowing the ambient.
If you know your ambient, hokies ambient, and your SB core temps you can estimate your eventual IB core temps using his percentile just fine. Alternatively you can simply look at the degrees over ambient (in C) and estimate that way. Both ways will work fine, neither will be exactly accurate unless you have the same cooler and fan of course.
Which will be closer? Probably *C over ambient.
Straight core temps are even worse than percentiles, degrees over ambient is the real measure of cooling.

Anyway, that's enough being pedantic for me.
The reality of the thing is that if people aren't looking at the hard data they probably don't know exactly what 46% higher means either, or don't care, so it is a moot point.

[jd101]

thanx. i have all new stuff ready and have a gene v z77 .just ant 2 sure to go with a 2700k or 3770k .

[EarthDog]

With a raised TJmax of 105C compared with 95C(?) for SB, I wouldnt think it would be that low...but have no clue personally.