• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Ci7 920 @ ~3.6 GHz and ~stock voltage = 90C

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

FuriousGeorge

Member
Joined
Mar 18, 2004
My RAM is running at ~1400 Mhz.

All my voltages are stock except vcore which is around 1.25, yet I can't break ~3.6 GHz (21x175) while keeping voltages under 90 C

Even with the case open I hit 90 C (as per RealTemp)

Ive tried two CPUs (Ci7 920 D0 and C1 stepping) and two MBs (Asus Rampage II Gene and ASRock Supercomputer). My HSF on one is that Noctuna (sp?) with 2 x 120mm fans, and the other has just the stock HSF.

Currently I'm using Noctuna thermal-paste, because I'm out of Arctic Silver, but I don't think that would cost 5 or 10 C in and of itself.

The HSF on the Rampage II with the D0, the open case, and the Noctuna is hot to the touch. It's around 80 C. That's an educated guess based on the fact that I know it can't be higher than 90 C, but meanwhile its hot enough that I can't continuously touch it with the skin on my knuckles (but I can take it on my finger tips).

That tells me heat is effectively being x-ferred from chip to HS.

I haven't tried every combination of Case, MB, CPU, HSF, and thermal paste possible, but I've tried most of them.

At first I had just the C1 stepping (though both MBs) with a mediocre thermaltake fan, and I thought I just unlucky...

Now that I have tried all those extra combos (especially the D0 stepping + Nuctuna + Open Case), I'm starting to think that maybe it's not the chip... maybe it's me...

Or maybe I'm just unlucky...

I hear about a lot of people who are able to pull off (~) 200 x 20 out of the box while barely adjusting the vcore. There's no way I could pull that off...

It's not that I care so much about the extra 300-500 Mhz, but I'm curious as to what might be the cause of these disappointing results...

Thoughts?

Thanks in advance!
 

nzaneb

Senior Member
Joined
Nov 3, 2008
My first suggestion would be directed towards your settings. Input your baseline voltages manually. Overclocking on Auto Voltage = Overvolting, on many motherboards.

Secondly (only suggesting as I'm not sure of your history... sorry), the TIM application could be excessive/ inadequate.

Thirdly... what stress testing program are you using? Linpack, Linx, Intel Burn Test will produce higher temps than say... Prime 95.

I would also suggest shutting off HyperThreading (for temp testing purposes) to see if your temps drop significantly.
 

Amtrak

Member
Joined
Jul 23, 2008
Yeah, if by stock you mean you left the voltage at auto, then the board is going to automatically pump in way more voltage than is often necessary, driving your temps up significantly. Use a program like CPU-Z while you're testing to see what your actual vCore is.

As also suggested, your TIM application could be problematic.
 
OP
F

FuriousGeorge

Member
Joined
Mar 18, 2004
I've pushed up to 200x19 / 1600 MHz RAM.

It's a bit hotter, but stable.

My vCore is 1.275 now
My QPI is also 1.275

My Dram voltage is 1.6, which may seem high, but my RAM is rated at 1.65 when @ 2000 Mhz.

I guess the only thing left to check is my thermal paste, then, because I didn't leave my voltages on AUTO.

Maybe I just got unlucky twice in a row.
 
OP
F

FuriousGeorge

Member
Joined
Mar 18, 2004
That was the temp on the last C1 stepping on the ASRock. I thought 50 C to the max was a good amount of headroom.

I guess I'll pop it open, pop off the HSF, reapply thermal paste, and see what happens.
 
OP
F

FuriousGeorge

Member
Joined
Mar 18, 2004
I was assuming he meant CPU Temp @ Ambient.

My ambient temperature is about 'room temperature'. Sometime a little higher, and sometimes a little lower. Let's say 19-21 degrees.

The good news is that reapplying the thermal paste seems to have lowered my idle temps around 8C (now 44-40), and after only a few minutes of running prime95 I notice a difference.

Load temps are maybe 12 C lower for the same amount of time running p95.

Looks like load temp is gonna max out in the low 80s to high 70s.

I got a lot of headroom with my voltage here, so I'm gonna try to push it a little higher, and check for stability.
 

CompuTamer

Member with Some Fancy Text Under His Name
Joined
Jan 4, 2009
Location
Brandon Mississippi
My chip pushes 90C under load too. 3.9 at 1.375 volts and it's pretty warm. Heatsink is too, so it's cooling... these things just seem to enjoy running hot (i'm an early adopter though)
 

DocClock aka MadClocker

Senior Member
Joined
Dec 19, 2000
Location
Stockton Cal, USA, Earth
I don't know about how good the Noctua TIM is, but I know for sure that good compound vs bad can make up to a ten degree difference in cpu temps...the 1st time I applied AS to my PIII, I saw a 7c drop in cpu temps.
Also, if your h.s is hot to the touch, then you need better fans (more CFM) or is it cubic inches per minute? At any rate, you need more air.
 

Conumdrum

Member
Joined
Sep 15, 2007
Location
Small town Emlenton, PA
Man, I would never push my i7 or i5 etc in load tests to 80C. If it loads at 73C I'm done, I'll play with my settings to drop temps and increase clocks, but 80C load under test? Not for me.

What is the max temps for an i7/i5 by the Intel specs? It's not 100C I can say that.
 

Conumdrum

Member
Joined
Sep 15, 2007
Location
Small town Emlenton, PA
Here ya go:


Once again, my objective is to assure that enthusiasts understand Intel's specifications, standards and test methods, so they can better decide how to apply and manage their overclocking options.

From Intel's Processor Spec Finder - http://processorfinder.intel.com/L [...] SearchKey=

All Core i7 9xx variants:

Vcore Max 1.375v
Tcase Max (CPU temperature) 68c
Tjunction (Core temperature) 73c

The Thermal Specification shown in Intel's Processor Spec Finder is Tcase Max (CPU) not Tjunction (Core), which is a very common misconception among most enthusiasts. Since there's a 5c gradient between the CPU sensor and the Core sensors, (shown in the following Intel document) - http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf - just add 5c to the value shown in the Spec Finder to determine the corresponding Core temperature, which is 73c for all Core i7 9xx variants.

Intel's second and frequently misunderstood Thermal Specification, Tjunction Max, (100c for all Core i7 9xx variants) applies to overtemp protection such as Throttle and Shutdown, so you don't toast your transistors. As such, any i7 9xx Core temperatures which exceed 73c should be considered "overtemp". Further, when specifications are exceeded, then processor degradation becomes a concern, which is explained in the following AnandTech article - http://anandtech.com/cpuchipsets/i [...] i=3251&p=6

"Comp" This guy knows what he's talking about

original link
http://www.tomshardware.com/forum/255263-29-temp-questions
 
OP
F

FuriousGeorge

Member
Joined
Mar 18, 2004
I don't know about how good the Noctua TIM is, but I know for sure that good compound vs bad can make up to a ten degree difference in cpu temps

I dunno either, but I figured that a company selling an expensive aftermarket HSF would include paste TIM that is at least almost as good as AS5


Conumdrum said:
Once again, my objective is to assure that enthusiasts understand Intel's specifications, standards and test methods, so they can better decide how to apply and manage their overclocking options.

From Intel's Processor Spec Finder

I've read that. I've also read this which basically says keep it under 80. Granted, one's a blog and the other is from the horse's mouth...

That said, I tend to take anything Intel says with a grain of salt. Intel also claims aftermarket cooling is grounds for voiding a warranty.

Also, I don't really worry so much about chip degredation. I've overclocked every CPU I've had for years, and I've never had one crap out.

Processors generally last 10 years, give or take. If mine dies twice as fast, then that means I've been using the same CPU for too long ;)
 

Neuromancer

Member
Joined
Oct 11, 2005
Location
Tau'ri
I do not use cases very often, but the NZXT zero, If I undid the thumbscrews and let the side panel pop out just a 1/4" CPU temps went up 3-5C because the airflow shifted just enough. (I do run balanced to negative case pressure designs though, positive just never works right for me)

Only take the side panel off your case and expect lower temps if you are directing a large fan at it... (or you have a really terrible air flow case) and not just straight at it.. force the air towards the leading edge of the motherboard so it gets pulled thorough the heatsinks (CPU/GPU) and expelled out the back of the case.
 

Bobnova

Senior Member
Joined
May 10, 2009
Tcase is the max temp the cpu will hit in 30*c ambients on the stock fan at stock voltage. It is NOT a limit, and has NOTHING to do with the danger zone or anything else, it is simply so that OEMs know how hot the CPU will get in certain conditions. Ignore it completely.
Tjunction is where the CPU throttles to avoid melting, it's the important one.
 

RGE

Senior Member
Joined
Feb 16, 2008
Location
East Coast
Tcase is the max temp the cpu will hit in 30*c ambients on the stock fan at stock voltage. It is NOT a limit, and has NOTHING to do with the danger zone or anything else, it is simply so that OEMs know how hot the CPU will get in certain conditions. Ignore it completely.
Tjunction is where the CPU throttles to avoid melting, it's the important one.

+1

btw, the 5C from core sensor to tcase sensor in article referenced above is on a 65nm chip, not an i7. Also, this 5C gradient is from core sensor (in die) to "cpu" sensor (still in die substrate) and has nothing to do with IHS temp or IHS spec max temp listed by intel.

For I7, theoretically if everything was stock, intel exact testing load conditions, as tjmax (100C) is reached on core, the IHS temp would be at max (68C at IHS), ie ~32C gradient, and cpu sensor would read closer to core temps, suppose was 5C and read 95C as cpu temp...completely meaningless number, nothing to do with core temps since it misses hot spots, and nothing to do with IHS or tcase max specs of 68C. Typically mobo bios calibrates the cpu sensor to read lower than its actual location to guess at IHS temp, but again very inaccurate, and useless on core i7 when you have accurate core sensors. Intel uses core temps for protection, because they are more accurate, and out of spec according to intel is "long time at or near throttling".

Hitting 100C is out of spec, running near 100C throttling temps for long time is out of specs. 80C is not out of spec, though may be over ones personal preference.

Granted for every 10-15C lower temps, mttf is halved, and nothing wrong with personal preference wanting lower temps, but I worry more about temps on things that do break like hard drives. My only obsession with temps on cpu is benching higher with better temps. Granted i wouldnt load 90+C folding at home 24/7, but nothing to base that on other than my personal preference.

btw...same argument here:
http://www.xtremesystems.org/forums/showthread.php?t=228788&page=2
 

Airbornederekc

Member
Joined
Mar 14, 2004
Location
Massachusetts
bump your qpi and uncore voltages to get a higher OC. the fact your CPU is running at 90c is most likely a bad mount or paste on the cooler.

What cooler are you using?