I7-3770K Temps?

[WilliamP]

Tweak Town ran overclocking tests on some CPU's including the Ivy Bridge I7-3770K and noted the high temp of the 3770K. I have seen this mentioned before and it has me confused and worried. I was planning a new build with a 3770K. First of all the Ivy Bridge is a lower wattage processor. In electricity, wattage equals heat. They did not indicate what cooler they were using. Hopefully it was the stock cooler.

 

[Smartazz]

It consumes less electricity than Sandy Bridge, but it's dissipated over a smaller die. You'll get less heat generated, but the die will run hotter. It'll be interesting to see if the engineering samples run hotter than the consumer chips.

 

[Kenmitch]

It consumes less electricity than Sandy Bridge, but it's dissipated over a smaller die.

Although the die shrink has it's advantages heat production isn't one of them. The cores are smaller and the heat is more concentrated which explains why Intel dropped the TDP.

Ivy Bridge overclocking will be dictated by temps most likely. If your going for a higher end overclock you'll need to choose your thermal solution wisely

 

[Smartazz]

Although the die shrink has it's advantages heat production isn't one of them. The cores are smaller and the heat is more concentrated which explains why Intel dropped the TDP.

Ivy Bridge overclocking will be dictated by temps most likely. If your going for a higher end overclock you'll need to choose your thermal solution wisely

Kind of funny that Sandy Bridge overclocks are typically limited by voltage whereas Ivy Bridge is limited by temperature. I still have a hard time believing that Ivy Bridge runs as hot as reviewers are stating.

 

[Kenmitch]

Kind of funny that Sandy Bridge overclocks are typically limited by voltage whereas Ivy Bridge is limited by temperature. I still have a hard time believing that Ivy Bridge runs as hot as reviewers are stating.

It's more sad than funny for all those who have been putting off an upgrade waiting for the latest and greatest thing.

SB's ramp up the heat very quickly under extreme loads but is maintainable if one has good cooling. The higher the overclock the better the cooling needed. Most of the time the cooling gives out before the chip is in the danger zone as far as vcore....Duds not included.

IB's look like they'll top SB's as far as ramping up temps. The maintainability of these temps is yet to be seen. Die shrink for the loss here it looks like!

Note to Intel....Ditch the damn heatspreaders and rethink cooling solutions! Seems like they could come up with a somewhat semi-idiot proof way if they wanted. I think the main thing is they don't see the need for it for the masses.

Maybe a new line for those who like to overclock the K's that doesn't have a heatspreader would be the best option for Intel. Seems like the danger of silicone damange could be minimalized by something like high heat rubber pads in all 4 corners of the heatsink....Just a thought

I think Intels goal is to get more power in a smaller place while using less power more than anything....Going green for the loss to overclockers!

 

[WilliamP]

I have a Turniq 120 Tower Extreme for cooling. I hope it will be enough to cool the 3770K.

 

[aaksheytalwar]

If you can get SB to run full load below 70c, you will be fine with similar clocks with ivy I guess but with 80-85+ c

 

[Diogenes2]

Kind of funny that Sandy Bridge overclocks are typically limited by voltage whereas Ivy Bridge is limited by temperature...

Where are you seeing Ivy Bridge overclocks being limited by temperature ?

 

[aaksheytalwar]

Reviews are out my friend. Check tweaktown

 

[Kenmitch]

Where are you seeing Ivy Bridge overclocks being limited by temperature ?

Smaller die = smaller cores = concentrated heat = higher temps.

All overclocks are dictated by temperature as well as voltage. Depending on a persons choice for a cooling solution they will hit the thermal wall before the chip gets into the vcore danger zone.

From this point here on I'm not attacking you. Just trying to get some reasoning for your current overclock is all. No harm no foul

Your current chip is running at 4.4ghz did your chip crap out at that? Did your cooling solution combined with your perception of livable temps dictate it? Was it the vcore that stopped you there? Just trying to get some insight is all.

 

[Fjodor2001]

Ok, but if you're not going to OC, and instead is interested in building a low temperature silent PC, IB should still be better than SB, right? The heat may be concentrated on a smaller chip area on IB, but the total heat will be less so it doesn't matter in that case, right?

 

[Smartazz]

Ok, but if you're not going to OC, and instead is interested in building a low temperature silent PC, IB should still be better than SB, right? The heat may be concentrated on a smaller chip area on IB, but the total heat will be less so it doesn't matter in that case, right?

If running stock, then yes Ivy Bridge will be better because it will consume less power. Ivy Bridge is probably going to be good for laptops where energy usage needs to be kept low. I'm curious to know how hot it runs with the stock cooler at stock speeds and in laptops.

 

[Imouto]

It will run hotter but will output less heat.

IB: 160 mm² SB: 216 mm² IGP: 12 EUs
IB: 1.4 b SB: 1.16 b IGP: 16 EUs

So I would assume that the IGP is close to 33% bigger than SB one. The overal 26% die size reduction would be even greater for the CPU having a very little area to dissipate the heat even after being lowered in the die shrink.

 

[Tempered81]

Tweaktown hit 4.8 on water. Bigtime suckage.

If you're overclocking for the 4.5-5.0 range, 2700k >3770k. Take into account the 8% (being generous) clock-for-clock advantage the 3770K has over the 2700K, it's a wash for max OC.

Chilled liquid, LN2, Dice, cascade, helium, w/e you can use the 64x multi. Air and normal water not looking good for > 48

 

[Imouto]

In fact since cache is the same I think that 20% more transistors are all packed in the new IGP since IPC isn't that much better leaving way less space for the actual CPU.

 

[Kenmitch]

Tweaktown hit 4.8 on water. Bigtime suckage.

If you're overclocking for the 4.5-5.0 range, 2700k >3770k. Take into account the 8% (being generous) clock-for-clock advantage the 3770K has over the 2700K, it's a wash for max OC.

Chilled liquid, LN2, Dice, cascade, helium, w/e you can use the 64x multi. Air and normal water not looking good for > 48

That's pretty crazy! Just looked at Tweaktowns preview.

Seems like die shrink + hyper-threading isn't a good combo as far as temps go.

The i5-3570k doen't look like it would be too hard to tame.

 

[Termie]

That's pretty crazy! Just looked at Tweaktowns preview.

Seems like die shrink + hyper-threading isn't a good combo as far as temps go.

The i5-3570k doen't look like it would be too hard to tame.

Link for those who are interested: http://www.tweaktown.com/articles/4...e_i7_3770k_and_core_i5_3570k_cpus/index8.html

 

[Smartazz]

Link for those who are interested: http://www.tweaktown.com/articles/4...e_i7_3770k_and_core_i5_3570k_cpus/index8.html

Interesting that voltages are on the low side, but temps are high, even under water: http://www.tweaktown.com/articles/4...e_i7_3770k_and_core_i5_3570k_cpus/index2.html

Edit: Oops, I forgot that CPU-Z doesn't read voltages on Gigabyte boards correctly.

I'm so glad I got impatient and went with a 2500K. My motherboard is compatible with Ivy Bridge, so if it turns out that early samples had something wrong, I can always upgrade. Unfortunate for anybody who waited if this is true.

 

[Diogenes2]

Smaller die = smaller cores = concentrated heat = higher temps.

All overclocks are dictated by temperature as well as voltage. Depending on a persons choice for a cooling solution they will hit the thermal wall before the chip gets into the vcore danger zone.

From this point here on I'm not attacking you. Just trying to get some reasoning for your current overclock is all. No harm no foul

Your current chip is running at 4.4ghz did your chip crap out at that? Did your cooling solution combined with your perception of livable temps dictate it? Was it the vcore that stopped you there? Just trying to get some insight is all.

It took 1.4v to get 4.8 stable, temps in 80's..
I don't worry about temps unless I'm crowding the throttle point..
Power from the wall was over 500w, and not what I care to run 24/7 .. ( Folding )
I can run 4.6 at 1.35v but the performance gain for me isn't worth the extra power required..

My point about questioning Ivy Bridge overclocks being limited by temperature, was that it simply hasn't been established yet.

But I will agree, it appears that IB may not be an attractive overclocker due to the temps being reported, but ultimately the overclocks will be limited by the ability and willingness to take the steps necessary to keep the temps under the throttle point.

I don't see it being worth the effort for most who have an SB at 4.5 or higher..

 

[Imouto]

 

[Kenmitch]

My point about questioning Ivy Bridge overclocks being limited by temperature, was that it simply hasn't been established yet.

I agree. The jury is still out on the temp issues. Only thing is the chatter being heard in the courtroom doens't look so good for the accused

 

[Lepton87]

Removing that stupid heat-trapping IHS should help a lot.

 

[Smartazz]

Keep in mind that the number of transistors dedicated to the GPU has likely gone up meaning that the actual CPU cores occupy a smaller space than that picture would otherwise suggest.

 

[Rvenger]

Interesting that voltages are on the low side, but temps are high, even under water: http://www.tweaktown.com/articles/4...e_i7_3770k_and_core_i5_3570k_cpus/index2.html

I'm so glad I got impatient and went with a 2500K. My motherboard is compatible with Ivy Bridge, so if it turns out that early samples had something wrong, I can always upgrade. Unfortunate for anybody who waited if this is true.

CPU-Z and Gigabyte boards misread the Vcore. Its reading the VCCIO I believe. There is something wrong with this review. How come the 3570k could only get to 4.6ghz?

"In the end, though we did end up with a slightly higher clock than the 3570k thanks to a 48x multiplier which as you can see above bought us in at 4.8GHz. These aren't huge overclocks, we've had our 2600k running at 5.2GHz and we wanted to ask around a bit to make sure we weren't doing anything wrong"


With the way things are looking, I might end up on X79 or with a 2700k

 

[Smartazz]

CPU-Z and Gigabyte boards misread the Vcore. Its reading the VCCIO I believe. There is something wrong with this review. How come the 3570k could only get to 4.6ghz?

"In the end, though we did end up with a slightly higher clock than the 3570k thanks to a 48x multiplier which as you can see above bought us in at 4.8GHz. These aren't huge overclocks, we've had our 2600k running at 5.2GHz and we wanted to ask around a bit to make sure we weren't doing anything wrong"


With the way things are looking, I might end up on X79 or with a 2700k

Oh yeah, I forgot about that issue. Anyone have any idea how accurate the CPU-Z voltage reading is normally? I typically see about 1.22v under full load and I'm curious how accurate this is.