That's my major concern, the space heater effect over time.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Post your New Ivy Bridge Processor Thermals, Type of Cooling, and First Impressions:
- Thread starter MeanBruce
- Start date
That's my major concern, the space heater effect over time.
In this instance the 60c chip will as it's the chip pulling and using more energy (95W vs 77W). As I've already said, this is a simple matter of energy conversion. The reduced die size of IB coupled with a less robust interface between the silicon and the heat spreader results in a less efficient rate of dissipation of stored heat within the CPU core, so yes the IB CPU is hotter when measured by the CPU's thermal sensor. However this does not change the fact that the power requirement of the SB processor is greater and as such a larger amount of energy is being converted to heat and being transferred away from the core, it's simple being done more efficiently thanks to the use of solder over TIM and the larger surface area of the die (the amount of die size shrinkage on IB is disproportionate to the reduction in wattage requirements, hence the higher concentration of heat).
It's the same basic principle as heating up a bucket of cold water using a thimble of 100 degree water vs a coffee cup full of 80 degree water, there is still a greater amount of energy in the 80 degree coffee cup and thus it will do a better job of raising the ambient temperature of the bucket.
Think of it this way: You take a cpu and run it with a stock cooler. Take the same cpu and stick on an aftermarket cooler. In the second case it will show a higher temperature, but the amount of heat put out into the room is the same. The IB chips also have a different internal way of dissipating heat so they tend to have higher core temps even though they use less power.
If you key concern is that of raising the ambient temperature of the room, then that is indeed the source of heat you should be paying attention to.
Thanks for the explanations -- I guess it's hard to argue against the hopefully true scientific methodology there. Maybe the room was feeling warmer than usual yesterday; either way I still believe returning the 3770k was right for me.
Exactly, so all these reviewers are just oc'ing and calling Ivy a thermal dud, simply because the nanometer cores run hotter? Who cares if your room runs much cooler?
I should have said, they are measured warmer. Maybe they don't run hotter, if dissipation is just poor at the nanometer level.
Last edited:
That still doesn't tell you anything, because case airflow plays a large part in that. A larger volume of cooler air can have a greater total effect than a smaller volume of hotter air (unless that air is blowing directly on or near you - which changes the situation somewhat). What you need to measure is the source of the heat - and that is the power used by the system. Ivy uses less power, so it is dumping less total heat into your environment. Period.
People who don't understand thermodynamics. However, the temp of the cores does have a negative effect on overclocking, so it can be a problem even if it is using less power. It depends what your concern is - if you want your room cooler, than you want Ivy, but if you want high overclocks, it isn't so great for that.
People are too wrapped up in the temperature even when it doesn't matter, like at stock. We see the same thing with video cards - some cards are allowed to run hotter so a cheaper or quieter cooling system can be used. As long as the system does what it is supposed to do, and doesn't exceed thermal tolerances, the actual temp of the chip doesn't really matter much over the kind of lifespan the usual entusiast is going to care about.
Last edited:
edit: I see this was already clarified.
So there is hope for Ivy, less heat pushed out of your chassis, less heat distributed into your working space. Explain further please.
So if all the proper laws of thermodynamics are still in place, Ivy is a success at stock speeds and overclocked? I'm doubting no one, just trying to understand this.
Last edited:
It depends what you are looking for. If you want high overclocks, then you want to keep the temps down because the chip will start to throttle at some point, and higher temps increase resistance, which means you have to push more and more volts in to keep going higher. So Ivy isn't great for high-end overclocking.
If all you want is moderate overclocking, say 4.4 or so, than Ivy is going to use 170W instead of 230W for the same overclock on Sandy (system power, and all numbers approximate and made up). So the Ivy chip, even though the core temperature is higher, is actually putting less total heat into your room. The temperature is not indicative of the power use - and that is what a lot of people seem to be missing in the discussion of Ivy's temperatures. A candle is a lot hotter than a hot-plate, but which one heats your coffee better?
However, if your room has even moderate airflow (an open window, say) the difference between the two probably isn't going to be noticeable, because the changing of the air in the room will dissipate and remove the extra heat anyway. It'll matter in a closed room a little, or it'll matter a lot if you have the computer in something like an AV unit (for an HTPC, say).
NathanP2007
[H]ard|Gawd
- Joined
- Dec 17, 2007
- Messages
- 1,697
I think the best way to close up this heat argument is: Measure the heat of the air blowing off your heat pipes from your tower cooler, or from your Corsair H fans on the radiator. Is the air they are blowing (the "Hot" air filling your room) really 20C degrees hotter than the air being blown by the same cooler on a SB cpu? Just because the sensor says the IB CPU is 20C higher (80C) at full load, compared to your SB CPU at full load (60C) doesnt mean anything per say about the temperature change of the actual air coming off the fans that are cooling your Tower/Radiator.
Give it another chance, you are still within your 30day window.
Sadly I already packed it up and sent it off to the UPS store earlier today to be returned to amazon. Even despite learning of the error in my thinking in terms of heat dissipation, it's probably the right thing for me to do anyway -- had no business "upgrading" from a 2600k @ 4.5GHz, hehe.
For extended use at stock speed 3.5Ghz, running office apps 8 to 10hours, in Florida and Summer is coming, my greatest concern is having to turning the air down for the entire house, just to keep my single room office from overheating. For this circumstance Ivy is a better choice over Sandy, even if only slight? Sorry so many questions, my 3770K arrived today and am really wanting it to be the right choice. Thanks so much.
If Ivy’s increased core temperatures do not equate to increased room temperatures, then what’s this thermal issue all about? Or is there even an issue?
Last edited:
Been playing around
1.225v at load. mostly hovers at 1.208v 
adding load pic in a minute... \\edit damn. froze that time... It's taking a lot of voltage above 3.6GHz. 3.6 seemed stable at 1.2v but I only ran 10 on high a couple of times before turning it up.
Last edited:
It is the right choice, yes. But at stock speeds the difference is really not going to be all that much - you are looking at 20-30W or so, so ultimately less than the difference in having another light turned on (even though it does add up over time).
The thermal issue is about two things - people getting bent out of shape that it runs hotter than SB, whether that matters to them or not, and the fact that it doesn't overclock as high as SB (which honestly may or may not be because of the heat issue).
I even installed LED lighting in the office just to cut down on the heat and am moving from a CCFL monitor to LED for the same reasons. Thanks so much for your help.
Moving from a Core2Duo and an X48 motherboard with a northbridge that runs so warm, the heatsink would give 1st degree burns. Hoping for more with Z77.
Last edited:
You are completely correct. There is really little reason to upgrade from a 2600k unless you have money to burn.
An IB could run hotter than a SB but give off less heat. You don't need a Phd in thermodynamics to understand this, it is common sense.
Think about it, when you take the the temperature on a chip, you are measuring heat at a single point, whereas when you talk about a unit giving off heat you are talking about the total heat produced by that item over its entire surface, and since the IB and SB have a different amount of surface areas, if they had the same measurable temperature at every point on their surface the SB would be giving off a lot more heat than the IB because it is bigger.
The fact that the IB has a smaller surface to dissipate heat is compensated somewhat by the lower power requirements of its .22 nm design, but not enough, apparently. That could be no problem, we just have to figure out how to more efficiently dissipate this extra heat from the smaller surface which we didnt have to worry about with the SB. So we use TIM instead of solder for the first and most important connection to the cooling system.
You are completely correct. There is really little reason to upgrade from a 2600k unless you have money to burn.
An IB could run hotter than a SB but give off less heat. You don't need a Phd in thermodynamics to understand this, it is common sense.
Think about it, when you take the the temperature on a chip, you are measuring heat at a single point, whereas when you talk about a unit giving off heat you are talking about the total heat produced by that item over its entire surface, and since the IB and SB have a different amount of surface areas, if they had the same measurable temperature at every point on their surface the SB would be giving off a lot more heat than the IB because it is bigger.
The fact that the IB has a smaller surface to dissipate heat is compensated somewhat by the lower power requirements of its .22 nm design, but not enough, apparently. That could be no problem, we just have to figure out how to more efficiently dissipate this extra heat from the smaller surface which we didnt have to worry about with the SB. So we use TIM instead of solder for the first and most important connection to the cooling system.
If a temperature reading at the 22nm level does not equate to an increased temperature reading in the room, I don't care if RealTemp says 100C.
Apparently this concept ranges between difficult and impossible for many people to understand. Seems many cannot differentiate between heat and temperature.
lol
Whats wrong with these pictures? Could the traditional temp monitoring programs (core temp/real temp) not be reading IB properly?
Load Temps:
Idle Temps:
Since I didn't see anyone address this.
Realtemp uses the sensors on your CPU, Asus uses the sensor on your motherboard. That's why the motherboard is not as hot. If you go just by your motherboard CPU temp, you're going to be awfully upset when your CPU dies later because it's pegged at tjMax and you think you've got a magical chip that is 20c cooler than everyone else.
PhD? I only have an MD just research and I cannot make sense out of this Ivy mess, too much Neurobiology, not enough Fluid Dynamics I guess.
I'm young, don't see patients, I write, no malpractice insurance, I love it. Why do I feel the need to share this? It's boring to others, hmmm. Signing out Hard.
Last edited:
That looks really good - what are you using to cool that chip?
The change in temperature in a room is a factor of how much heat is being generated by the chip and how much of this heat can be grabbed/dissipated through the cooling system into the air in the room.
You probably can measure how efficient your cooling system by figuring how much total heat the chip is producing (there now you can consult your thermodynamics professor about this), and compare it to the heat being generated on across the entire surface of the cooling system (mostly the radiator). Ideally, you would want the chip and the radiator surfaces to be producing the same amount of heat, but I dont think this is a realistic expectation in an air cooled system.
I would get worried if my radiator was not hot enough, that would likely mean it is not doing its job. BTW Bravo for Tom and the Inquirer: http://www.tomshardware.com/news/ivy-bridge-overclocking-high-temp,15512.html
Just a Cooler Master 212+. Ambient temp is about 75F in my room right now. That was right between runs I guess, notice the max temps, those are just a bit above where it usually sits at load.
That's a nice result, 4.7 @ 80C is what a lot of water cooling guys were reporting (although those may have been 3770Ks). Maybe another reason to look at 3570Ks, if they run significantly cooler.
Yeah, it seems to me the heatsink just doesn't make much of a difference with these chips. The 212+ is noticeably cooler to the touch with this chip than it was with my x6.
It's ok, unfortunately some people like to get their rib in for what seems like no other reason but to gain some sense of false superiority complex (small pen__ syndrome?)...who knows, who cares really. At least something new is learned.
SonDa5
Supreme [H]ardness
- Joined
- Aug 20, 2008
- Messages
- 7,437
Actual surface area of die of 22nm IB may be so small that is matters very little if water cooling or air cooling is used at ambient temperature.
The magic of IB for extreme over clocking is happening with sub ambient coolant surfaces on the die.
I'm hoping to hit a stable 5GHZ on my 3570k with my water cooling loop.
One thing that I have noticed is over all temps on my water loop are cooler with IB that with SB and this helps my video card run a little cooler as well.
The IB chip does get hot but it doesn't heat up my loop as much as SB even though it ran cooler. (Good thing for water cooling when loop cools more than 1 block).
Wow.... excellent news... I was planning on picking up 3570K and use my 212+ from my current Phenom II build (see my sig). How is the system feel comparing to your Phenom II build? Also what theremal paste have you used for 212+? Are you using 2 fan configuration on your 212+? What technique have you used to apply paste? I'm sorry for all the questions... and thanks in advance for your help
Yeah, seems the difference between half way decent cooling and great cooling isn't being shown in the core temp of IB. My temps are virtually the same when using a single fan vs push/pull config. The heatsinks are not at all getting heat soaked, the heat simply isn't being dissipated fast enough. The greater the temperature difference between the core and the cooling apparatus, the faster the heat transfer which is why the extreme cooling solutions work so well.
Clearly didn't bother reading past that post, It was addressed multiple times almost immediately.
I wasn't thinking my chip was magical, I was inquiring as to the possibility that the traditional temp monitoring programs may be reading IB temps incorrectly.