Ivy Bridge Heat Problem

itxmini

Honorable
Mar 11, 2012
12
0
10,510
I have recently heard an issue having to do with unsuspectingly or even outrageous temps with the new Ivy Bridge CPUs.

I was wondering if this might be one of intel's faulty or defective batches.

I have been waiting a ling time for their release of Ivy Bridge processors and was wondering if i should just go ahead and buy the SB i5-2500K ( going to be overclocking) or should i wait if intel can confirmed and fixed this heat issue.
 

xtreme5

Distinguished
may be thats will be one the defective chip not all the ivy bridge cpu's if so then why intel develop thier product which is faulty so thats not the case wait for ivy.

and from where you heard this link me. ivy bridge is not available yet so how is this possible.
 

catatafish

Distinguished
Feb 6, 2012
448
0
18,810



Start here: http://www.tomshardware.com/reviews/ivy-bridge-benchmark-core-i7-3770k,3181.html
 

itxmini

Honorable
Mar 11, 2012
12
0
10,510


the new ivy bridge CPUs were officially announced today. Many reviews by well known websites regarding the i7-3770K and i5-2570K have been posted. just do a quick google search and you will see.

anyway, these new processors (not just one, many tested by many different people and reviewers) have been tested with even temperatures of more than 10-20C compared to Sandy bridge when overclocking even at low cpu voltages and adequate cooling such Corsiar H100.
 

catatafish

Distinguished
Feb 6, 2012
448
0
18,810


If you're uncomfortable due to rumors, just buy the sandy bridge. As the article I posted mentions, there aren't really epic performance advantages to be gained. This is Intel moving slightly forward with performance, while transitioning to a smaller die.

Personally, I don't put a lot of weight in internet posting people claiming what their temps are in completely non-scientific testing. So far, I've seen compu-journalists, who control for variables far more objectively, claim a 10c difference, holding other variables constant. At my overclock, I could live with 10c easily. If you're going to overclocking extremes, then sandybridge is probabably better for you, but probably too soon to tell.
 

MKBL

Splendid
Nov 17, 2011
429
3
24,565
All the reviews on Ivy Bridge boil down to how wonderful this chip will be for laptop. For desktop department, they are less enthusiastic, unless they emphasize on videocoding. Also, they showed less stable overclock in Ivy Bridge. But with the same pricing or better, Ivy Bridge is a winner for those who don't care about those.

Someone succinctly put it this way: If you are overclocker, go for Sandy Bridge. If not, Ivy Bridge.
 

sburck

Honorable
Mar 28, 2012
30
0
10,530
As an overclocking newb, I had some similar concerns/questions. The articles for the most part say that ivy bridge does get hot when it's oc'd but it is stable. Is there a problem with the heat if the chip seems to be running fine at 4.7-4.8 ghz?
 
People need to learn to wait when new products hits the shelves. The problem isn't the die it is the crappy ihs that intel uses. I am waiting to see if anyone has the guts to try to pull the ihs off one who knows greedtell might have only used cheap gray goop instead of metal for thermal interface.
 


http://www.tweaktown.com/articles/4621/intel_ivy_bridge_overclocking_with_the_core_i7_3770k_and_core_i5_3570k_cpus/index8.html

95C in the graph OC'd to 4.7 .... and that's throttled down from a peak of 105C

They got a new sample and tested here to 75 ... but kinda puny OC

http://www.tweaktown.com/reviews/4673/intel_core_i7_3770k_lga_1155_ivy_bridge_cpu_review/index11.html



 


Wow that sucks, I would have expected more from such a cpu with the latest process. Crappy ihs or they made changes that limit clocks compared to SB architecture wise.
 
G

Guest

Guest

Thermal paste? Don't they solder it on?
 
Thermal paste? Don't they solder it on?

Low grade compound on low end models (that has never changed with intel and amd) and metal base on higher end premium parts. This being a premium part there should be a metal base instead of compound but the poor thermals suggest that there is indeed something not right going on. Either they didn't use gold electroplate with metal base or they used poor quality and uneven ihs on their samples. The small die shouldn't make cooling issues but there are limits to how much heat can be moved for a given mass for a given area. Either way they must have skipped something or messed things up. 77w tdp should mean cooler for a given voltage at any clock compared to higher wattage models.

So what I want to know is what they used or what they did wrong compared to SB. Low end parts that didn't use a metal base for thermal interface typically have this cooling related issue. Hell I tested it on a e7200 to see if changing the compound they used under the ihs with something better and max oc improved by 200mhz with temps slightly lower.
 
How far do you plan to overclock? From Anandtech:

"Those wishing for 4.8GHz at 1.4 volts (similar to Sandy Bridge) will run into a lot of issues if they think that 1.4 volts is appropriate for Ivy Bridge. In comparison, you may end up with something more reasonable like 4.6GHz at 1.1 volts, or 4.8GHz at 1.2 volts (as per some boards I have tested). Then it will be a case of deciding whether the small IPC gains that Ivy brings will be worth 200 MHz less on your CPU compared to Sandy Bridge."

Would you be happy with a 4.6GHz quad core? I will be - in about two weeks :)
 

gmkos

Distinguished
Jul 28, 2011
225
0
18,760


While I'm not too fond of voiding the warranty on a brand new CPU, I've been waiting too long for an upgrade and there is no way I'm getting a Sandy Bridge at this stage. Given the initial thermal results from vendors OCing test samples, I'm very much considering lapping an Ivy Bridge chip next week.
 

InvalidError

Titan
Moderator

The same things that should have made IB a better overclocker in theory also contribute to hinder it.

While multi-gate helps increase switching speed and reduce leakage, the higher proportion of surface area between gates and channels counters a large chunk of the reduction from the process shrink so it shouldn't be much of a surprise that dynamic power is ramping up faster than it used to with single-gate CMOS. Faster gates also means faster edge rates which adds its own share of engineering challenges for power delivery and decoupling across the core and LGA substrate.
 
G

Guest

Guest


I applaud you!
Hopefully they'll come out with benchmarks and see how much improvement we can get!