Post your New Ivy Bridge Processor Thermals, Type of Cooling, and First Impressions:

I am talking my time... just because I have to backup my current computer :(

But I did build my system (just need to take my videocard, ssd/hd, ram, etc out of my current system

523114_10100656372960417_29622837_50829105_693581310_n.jpg

I love it, the Noctuas go so well with the Sabertooth, good lighting. You are doing a really great job, I was either going to go ROG red/black or Sabertooth/Noctua earth tones, both good choices.

This is the Noctua Tri-Tower successor to the NH-D14, Dephcon was thinking of, pretty amazing.

p1000559.jpg
 
Last edited:
I'm leaving it on Auto and using "vdroop" settings in MSI bios to "offset". I haven't seen an "offset" setting. So I'm assuming it's my only option.

Here are my final settings for 4.5ghz.

This gives me all the bells and whistles + 4.5Ghz. I'm satisfied.

You can't change CPU Core Voltage to offset on that board? I know some of the lower end boards don't allow offset, but I'm surprised the GD65 doesn't.
 
I'm having to use negative offset since setting the voltage to auto raises it with clock speeds higher than it really needs.

Currently testing at 4.2GHz 1.160v under load according to CPUz.

LLC set to auto which as far as I can tell is using intel's spec.
 
I re-oriented the 212+ vertically so the fan is now blowing the air towards the back of the case. Same temps (70c w/p95 @stock). I'm using Arctic Alumina "premium ceramic polysynthetic thermal compound Arctic Silver". Maybe this grease is causing this issue?

Could be that. I don't think Arctic Alumina is particularly highly rated. Could be a few different things. Have you looked closely at the IHS for flatness? Bowed or caved in any way? How about the heatsink? When you take it off the IHS what does the paste look like? Any voids or areas where it doesn't look like there is good contact? Also, because Intel didn't use solder, I don't think you can rule out that maybe the chip itself isn't making good contact with the IHS. If you try a new paste and varying thickness and still can't get the temps down I would RMA the CPU because it seems others aren't having the same issues.
 
I doubt you'll be able to RMA the CPU unless you fib about why since those temps are still well within what intel would consider normal.
 
Man I am hoping for 4.6 @ 1.2V so far 4.5 @ 1.240 is holding great @ 59-61C

That would be good yes? What is a good volt / GHz to shoot for?
 
Well my GPU is blowing the hot our out the back of the case when in use, and in these tests I'm not utilizing it. It's not putting out any heat for desktop CPU stress testing here. The gtx680 is has very little idle heat production.

I re-oriented the 212+ vertically so the fan is now blowing the air towards the back of the case. Same temps (70c w/p95 @stock). I'm using Arctic Alumina "premium ceramic polysynthetic thermal compound Arctic Silver". Maybe this grease is causing this issue?

Even if your card blows most of the heat out the back there is still going to be heat coming off the top side of the card and since heat rises it will hit the fins of the cpu heat sink.
 
SO I am sticking with 4.3GHz at 1.19V Seems pretty rock solid and temps are at 55-60 under heavy load. Time to start loading up games and really see how it does.

Core Temp is telling different voltage though 1.26V but ASUS utility says 1.190 so I am guessing its good to go.

Not sure if I got a "Good" chip or not as I dont know all the new things with OCing but I did see 4.6 1.39V and then said no and stopped it.
 
I have some very interesting findings regarding temps... Been stress testing for over an hour straight now, RealTemp has hit a max of 72-79C for the seperate cores... Asus monitor is reading 57C? I'd LOOOOVE to believe Asus monitor
 
I got 4.5ghz dialed in cpuz says what my bios says voltage wise 1.2v
temps 75c prime95.

I would say about average chip.
 
Could be that. I don't think Arctic Alumina is particularly highly rated. Could be a few different things. Have you looked closely at the IHS for flatness? Bowed or caved in any way? How about the heatsink? When you take it off the IHS what does the paste look like? Any voids or areas where it doesn't look like there is good contact? Also, because Intel didn't use solder, I don't think you can rule out that maybe the chip itself isn't making good contact with the IHS. If you try a new paste and varying thickness and still can't get the temps down I would RMA the CPU because it seems others aren't having the same issues.

Both the IHS and heatsink looked flat against each other. When I reoriented it, there were no areas of varying thickness on either. Looks like a flat contact on the entire IHS. The user comments on Newegg for this grease seemed very good. *shrug*. Was only $5 though. I have some four year old Arctic Silver 5 I could use. Is that a bad idea to use old TG?

The 2700K I had installed in the same setup earlier today had some heat getting to the heatsink fins (but not a lot) as it was warm. The 3770K I can barely feel any heat on the heatsink. It's very subtle. Just a little warm nearest the base, at full q95 load. Seems to me like the CPU heat isn't getting to the heatsink, which supports the notion that perhaps the grease between the silicon and IHS isn't doing a good job.

BUT this is only my findings and there's a possibility I did something wrong so don't take my findings as a final report on IB temps, etc.
 
Last edited:
Whats wrong with these pictures? Could the traditional temp monitoring programs (core temp/real temp) not be reading IB properly?

Load Temps:
2jetw8n.jpg


Idle Temps:
96jnnl.jpg
 
Even if your card blows most of the heat out the back there is still going to be heat coming off the top side of the card and since heat rises it will hit the fins of the cpu heat sink.

FYI I have case door open. The CPU heatsink fins are close to room temperature. They don't feel warm at all. The vid card effectively blows gpu heat out the back of the case. Really, it's not a factor in this case (pun). It's (evga GTX680 SC) also not hot on the top side of the card either, unlike my previous ATI cards.
 
Whats wrong with these pictures? Could the traditional temp monitoring programs (core temp/real temp) not be reading IB properly?

Load Temps:
2jetw8n.jpg


Idle Temps:
96jnnl.jpg
Interesting! How funny would it be if all this time it's the programs that are not reading the IB CPU'S correctly.
 
Exactly! We need to figure this out lol. I posted in the anandtech forums as well. Lets see what the collective minds can come up with. *crossing fingers*
 
Whats wrong with these pictures? Could the traditional temp monitoring programs (core temp/real temp) not be reading IB properly?

What does RealTemp have for the Tjunction? Should be 105C, so that could be throwing off the temp readings. I'd be more inclined to believe RealTemp (or CoreTemp) than I would the Asus tool though. CoreTemp and RealTemp read the data straight off the CPU - who knows what the Asus tool is reading. So as long as Tjunction is set correctly, RealTemp and CoreTemp should be accurate.

Edit: Yay for math - obviously it is set to 105C, so I'd believe the RealTemp values.
 
What does RealTemp have for the Tjunction? Should be 105C, so that could be throwing off the temp readings. I'd be more inclined to believe RealTemp (or CoreTemp) than I would the Asus tool though.

It's at 105. Normally I would to, but in this case, with this being the tool designed for this board and these CPU's being as new as they are, I'm not so sure. Then there's the experience factor. I've touched many heatsinks/heatpipes while they were under load. The heatpipes, even at the base, don't feel anything like they are dissipating anywhere near ~80C. It "should" feel like a video card heatsink to the touch, but it doesn't. It, in-fact, feels just like every other CPU I have that runs between 50-65C under load.

If everyone can run their monitoring apps that came with their boards and compare with real/core temp maybe we can get a little more definitive.
 
Then there's the experience factor.

Unfortunately experience also says that, if one of those two programs is going to be wrong, it's going to be the Asus one. Hopefully this is the exception though.

Here's unclewebb's comment on Ivy from the forum:

I have read some forums where all sorts of conclusions are being drawn on how hot Ivy Bridge runs. That's nonsense. These conclusions are based on temperature data that can not be relied on for any of Intel's recent CPUs. Temperature data from Intel CPUs should not be compared to other CPUs in the same CPU family let alone comparisons to CPUs from different families.

RealTemp is a good program and the sensors are useful enough so that if you swap a heatsink you can see if things are better or worse but beyond that, Intel's core temperature sensors are not accurate enough or documented well enough to provide users with 100% accurate core temperatures.
 
Last edited:
Maybe, but again, this is a unique scenario. I'm not saying ASUS is correct, what I am saying is that If I get an IR gun and point it at the base of the heatsink, I can assure you it will read very similarily to any other CPU (besides IB) where core/real temp record between 50-65C. There's only two reasons for that.

1) The heatsink can't extract the heat fast enough and much of it is trapped
2) Asus is correct.

Im certainly willing to accept either one, it would just be nice to know which it is.
 
Maybe, but again, this is a unique scenario. I'm not saying ASUS is correct, what I am saying is that If I get an IR gun and point it at the base of the heatsink, I can assure you it will read very similarily to any other CPU (besides IB) where core/real temp record between 50-65C. There's only two reasons for that.

1) The heatsink can't extract the heat fast enough and much of it is trapped
2) Asus is correct.

Im certainly willing to accept either one, it would just be nice to know which it is.

Or option 3, which is that the temperature sensor on Ivy is not in the same location as Sandy, and is possibly located in a hotter part of the chip. So the chip as a whole isn't hotter, but the sensor is in a local hot spot. Or it's just a crappier temp sensor and even less accurate than usual. They are only designed to be accurate at temp close to Tjunction anyway.

Testing with an IR gun is not going to tell anything about Ivy - you might be able to get enough data on SB to draw some correlation between heatsink temp and core temp (but I doubt that too), but it wouldn't be valid to compare that across generations.
 
Yikes, those are some pretty terrible overclocks and heat.

Why not just got with SB and get 4.8Ghz - 5.0Ghz?

Who cares about voltage. You're talking a handful of dollars for the entire year lol. Literally under $10 dollars.

People coming from 920's I can understand.

But if you came from a 2500 / 2600k, oh jesus ........
 
Or option 3, which is that the temperature sensor on Ivy is not in the same location as Sandy, and is possibly located in a hotter part of the chip. So the chip as a whole isn't hotter, but the sensor is in a local hot spot. Or it's just a crappier temp sensor and even less accurate than usual. They are only designed to be accurate at temp close to Tjunction anyway.

If that were true then the Tjmax would have been adjusted accordingly. AMD places their sensors in a different spot but because of that their chips have a lower TJmax (60-70C) which explains the difference. Technically they don't run cooler, they just report cooler temperatures and have lower ceilings. Ivy on the other hand has only a 5C ceiling increase which can't possibly account for the 20C+ difference when compared to SB.
 
4.5Ghz @ 1.189v/1.82vPLL was good for 8 hours of prime 95. Max temp with my fans at low low speed case closed and all was 80c (1 core, others hit 75 max).

One thing that impresses me is how cool the air coming out the back of my heatsink is. I can definitely tell there are less total watts to dissipate. Just seems hard sucking them up from the cores.

You can't change CPU Core Voltage to offset on that board? I know some of the lower end boards don't allow offset, but I'm surprised the GD65 doesn't.

I know my previous asus board had vcore offset. I can't find a similar option on my Z77A-GD65, I'll ask in the MSI support forum when I have a chance. I'm "ok" for now with no offsets, but If I want to aim for 4.6/4.7 I'll need a larger offset to keep dynamic vcore going.
 
Hmm I can adjust voltages just fine.

They did change the name for the compensation thing. went from having "levels" to being +xx%.

By looking at your screen shots for CPU voltage it says "Auto", opposed to "[Auto]" which means it a modifiable value.
 
Yeah I never had any issues adjusting voltages. I was just saying that my updated bios seemed to work fine, while yours didn't. Which I find odd.
 
It's at 105. Normally I would to, but in this case, with this being the tool designed for this board and these CPU's being as new as they are, I'm not so sure. Then there's the experience factor. I've touched many heatsinks/heatpipes while they were under load. The heatpipes, even at the base, don't feel anything like they are dissipating anywhere near ~80C. It "should" feel like a video card heatsink to the touch, but it doesn't. It, in-fact, feels just like every other CPU I have that runs between 50-65C under load.

If everyone can run their monitoring apps that came with their boards and compare with real/core temp maybe we can get a little more definitive.


Whats wrong is there is 15% load in the one pic, 100% in the other. ;)

I guess I've been in the Ivy threads because I really did want one and look forward to one, but I personally see nothing special now. I do want these guys to enjoy their chips though so I'm going to back off and refrain from furtherp posting unless it's me showing my Ivy results.
 
Ok so I went back to My 2600k and Sold the Ivy ..I want OC the thing to death but you cant go past 4.8 to 5ghz with out that Figgin thermal Celling!. UGHHHHHH.... Its Not the Mobos Its the Figgin Chips. I SO WANTED THIS TO BE AWSOME. but its just a Smaller Sandy. so Ill Wait for the Tock of Intels Clock. :/
 
say I want to hit 4.0 and thats it, i dont mind going any higher.
would the i7 3770 non-K be able to hit 4.0ghz?
 
Given that direct HS contact with the die doesn't help much, the best "fix" for this may be the arrival of IB parts with even larger caches, equivalent to the current SB socket 2011 parts. The larger cache will make the die physically larger, providing more contact area between the die and IHS.

IB i7-3770K seems to have about 1/3rd greater die-to-IHS thermal resistance than SB i7-2700K (0.41 versus 0.28 degrees C/W, http://hardforum.com/showthread.php?p=1038667049#post1038667049). It's possible the TIM plays a role in this, but it's certainly true that the smaller die does too. Heat transfer scales with contact area, so even a soldered-to-IHS IB would have a higher die-to-IHS thermal resistance than the comparable SB part. The 4-core 8MB i7 SB parts are 216mm2, the i7 IB parts are 160mm2. All else being equal (e.g. if both parts were using solder), the area difference alone would make IB die-to-IHS thermal resistance 35% larger. It looks like it is actually 45% higher, so maybe the TIM accounts for the other 10%.

In contrast, a 4 core 10MB SB LGA2011 part is 294mm2 -- 35% larger than the 8MB part, in part from the larger cache and in part from all the extra I/O signals -- according to Wikipedia. If the area scaling holds for IB -- and it might not, IB LGA2011 area might be I/O limited or the IB caches may be denser -- a 4-core 10MB LGA2011 IB part would have an area comparable to a i7-2700K, and might have a similar die-to-IHS thermal resistance.
So it might be easier to overclock.
 
Whats wrong is there is 15% load in the one pic, 100% in the other. ;)

I guess I've been in the Ivy threads because I really did want one and look forward to one, but I personally see nothing special now. I do want these guys to enjoy their chips though so I'm going to back off and refrain from furtherp posting unless it's me showing my Ivy results.

That isnt the problem. Look carefully. There are two apps running in each pic. One is comparing temperature discrepancy under load the other while idle.
 
Well Realtemps look pretty accurate. Mobo utilities are notorious for being generic.
 
New i7 3770k here with Asus Sabertooth Z77 board. Spent all night trying to get it dialed in. Horrible temps. AS5 applied and re-applied twice. Using Corsair H100 with Push / Pull fans. Excellent contact with the CPU.

Used the ASUS Z77 UEFI Tuning Guide for overclocking, found here: http://www.mediafire.com/file/l34d6j...ning_Guide.pdf plus my vast exp with my Asus P8P67 Deluxe and 2600k.

Best overclock I can get that's stable is 4.2 and at full throttle my temps are, well, up there. 75 - 80c.

Was a huge mistake to invest the money in this project. Should have stuck with the 2600k at 5GHz and just bought a 2nd HD 7970. I feel stupid for not paying heed to everyone's advice.

SB at 4.4Ghz is about the same as IB at 4.2Ghz.
 
Back
Top