TDP of CPU vs TDP of cooler, is there a good ratio?

D

digicat

Limp Gawd
Joined
Aug 13, 2004
Messages
184
If a CPU has a TDP rating of, say, 91 watts, would a cooler that's rated for 170 watts be better than one that's rated for 140 watts, or does it matter?

Is there a ratio of CPU TDP vs Cooler TDP for maximum heat dissipation?
 
A cooler rated at 170 watts will keep the CPU cooler than the 140 watts one. Technically, all you need is a 91 TDP rated cooler to cool at 91 watt CPU. There is no ratio, there is only how cool you want your CPU to be and at what noise levels.
 
Does the TDP rating on the cooler usually assume max fan speed?

If you were to replace the fan on the 180 watt cooler with one that moved 40% less air, would that make the lower the TDP by 40% too?
 
lol you would have to give better info than this, cause some fans can push more air flow at slower rpms, and the air flow in the case, and the size of the cooler. I think the larger more efficient coolers dissipate heat with less air flow.
 
If there are cooler standards they tend to be by individual test sites "for comparison purposes only."
Another words testing may refer to open air or a case nothing like yours.
As pointed out some fans move air better at slower speeds.
Coolers are often built to a price point and appearance.
Shop for the best rated cooler in your price range, that mounts reasonably well. There is a reason the best coolers are large.
Your case airflow is also part of the equation. Just saying..
 
The set up will be like this:

i7-6700k, stock speeds on up. One GTX 980 ti video card.

Fractal Define R5 case, with it's 2 140 mm fans (68.4 CFM each). Default positioning is 1 for exhaust and 1 for intake.

The cooler I was going to try is the Cooler Master Hyper 212 EVO. It's rated for 180w. The 120 mm fan that's included provides between 24.9 and 82.9 cfm, depending on what speed it's spinning at. If it proves too noisy, I have a 120mm Noctua fan that pushes around 63 cfm.

It's 25% less air, so it seems like it'd cool less than the CM fan, but how significantly less?

Lots of numbers. CPU's rated at 95w. The EVO is rated at almost twice that. Does that mean that I have some flexibility to swap in quieter fans?
 
You are far too concerned with power ratings. The power rating are assigned by the manufacturer and do little for comparison between models. Plenty of sites review CPU coolers with high wattage CPUs and report the results. The Hyper 212 is good for the price, however it isn't a top tier cooler. For stock speeds you can use damn near any modern cooler without an issue to keep the CPU under 95C. For overclocking, it really depends on how hot you want to allow the chip to get. Most people try to have their maximum temperature be well under what Intel allows the CPUs to run at before throttling.
 
Last edited:
What Ocellaris said. You're going about this the wrong way. Check cooler reviews, preferably ones using your CPU or similar CPUs (relative performance of coolers will change depending on the CPU). Similar CPUs would be any of Intel's small die CPUs that use TIM. (Ivy Bridge and Haswell quad-core mainstream CPUs).
 
I think you would be "ok" making that choice. You could also put 2 fans running at slow rpms as well not to mention it doesn't cost a lot in case you choose to use it as a backup cooler. If theres zero chance of trading it out latter i might invest a tad more
 
digicat said:
If a CPU has a TDP rating of, say, 91 watts, would a cooler that's rated for 170 watts be better than one that's rated for 140 watts, or does it matter?

Is there a ratio of CPU TDP vs Cooler TDP for maximum heat dissipation?
Click to expand...

I wouldn't put much stock into manufacture ratings. I would look at reviews and base ym choice from there. Check out hardforum techpowerup, guru, and so on for good coolers.
 
Tsumi said:
A cooler rated at 170 watts will keep the CPU cooler than the 140 watts one...
Click to expand...
Not necessarily. Imagine the 140W cooler has 4 heatpipes and a huge fan, while the 170W cooler has 5 heatpipes and a wimpy fan. All else being equal, the 140W cooler will perform better at lower power, up until its heatpipes are overloaded.
 
Mr Evil said:
Not necessarily. Imagine the 140W cooler has 4 heatpipes and a huge fan, while the 170W cooler has 5 heatpipes and a wimpy fan. All else being equal, the 140W cooler will perform better at lower power, up until its heatpipes are overloaded.
Click to expand...

Not the case at all, if both coolers were operating at the maximum included fan speed. TDP rating ignores noise, and ignores any fan profiles the motherboard might utilize. It also ignores any alternative fans that may be used. Adding fan profiles and alternative fans into the equation is outside the scope of comparing factory provided TDP numbers.
 
Tsumi said:
Not the case at all, if both coolers were operating at the maximum included fan speed. TDP rating ignores noise, and ignores any fan profiles the motherboard might utilize. It also ignores any alternative fans that may be used. Adding fan profiles and alternative fans into the equation is outside the scope of comparing factory provided TDP numbers.
Click to expand...
The power rating of a heatsink is affected by the non-linear thermal resistance of heatpipes. They have low thermal resistance up until a certain power, then it rises. Adding more heatpipes increases the power limit, but designs with more heapipes don't always have lower thermal resistance at a given power level. That certainly does depend on the fan, as well as the number and geometry of the fins, and various other factors.
 
Trouble with comparing Hyper 212 EVO is it is rare to see on test sites.
Your case offers good cooling. What overclock you are looking for pretty much defines the cooler needed.
 
I'm looking to OC it as much as I can while keeping it quiet. I'm going to us a Fractal Refine R5 case, which I know has some sound dampening features. I figured I'd establish a baseline to start at stock speeds, then start moving it up and seeing what my ears can tolerate.
 
If you want a max OC and quiet, my suggestion is to consider liquid cooling. A couple of triple rads outfitted with low rpm fans can dump a lot of heat silently. Besides better temps and and a higher stable OC, the low noise is a major reson I have used liquid cooling for several years.
 
I'm not quite sure how to explain all of this. There is a lot of miss-information here.

CPU with 100 TDP can generate up to 100 TDP at stock settings.

A 100 TDP cooler by definition has the abiltiy to cool up to 100 TDP of heat.

A 200 TDP cooler by definition has the abiltiy to cool up to 200 TDP of heat.

That is the simple part,
How the CPU and cooler were tested has many loopholes.
It's all about the lack of a stringent scientific standard test procedure.

But even as lax as this standard is, testing and review sites are many times worse. Many of these testers and reviewers have little or no idea how to do an accurate test of of any sort, even a simple TP test.

Sorry for being crude, but stop and think a minute. We see all kinds of "cooler tests and comparisons" that are done in a complete system including a case (case us usually running stock fans) using room ambient air temp checked before and after testing with reviewer say this kind of testing simulatges "real world use".
That is total load of carp!!
1st, how many of us have the exact same system the tester/reviewer has? Anyone have the same system? Nobody?
Okaiy, so much for the 'real world' part. We now see the only 'real world' involved is the tester system.

2nd, using room ambient air temp is total carp too. Even with open bench testing the air temperature around bench is not always the same as rest of room. The critical temperature is the air going into the cooler. In a case this can be anywhere from a couple degrees to 20c warmer than room depending on cooler (different cooler fans move different amounts of air, but case move the same with each cooler) If case is moving more air than cooler, air going into cooler will not be as hot as if cooler is moving more air than case and cooler has to re-use it's own heated air .. making the air going into cooler higher and higher above room air .. and guess what CPU temps go up very close to same amount as cooling intake air temps do!

But back to the testers and their systems. They are not testing cooler performance. They are testing their system performance. Why? Because changing coolers in a system using room air temp only give us the CPU temp of that system with with each cooler tested. We only know how that system is performing with each of the coolers We have no idea how each of the cooler is performing. We can say cooler 'A' performed better than cooler 'B' in that system, but we have no idea how cooler 'A' and cooler 'B' will perform in a different system .. with diferent case airflow, different fan placement, etc.

To test coolers we at least have to know what the air temp going into the cooler is at all times
With the air and CPU temperautes take once every second we can see how much heat the cooler is able to remove from the CPU. But we must have both of these temps!!

But looking at thermometer in the room tos ee how warm the air is going into the cooler is like cooking Xmas dinner in the kitchen and going into a bedroom to look at thermometer to see how warm the air in the kitchen is.

That make any sense? I mean, are you following my logic, not the testers illogic. :D

I hope this makes sense to at least some of you.
 
  • Like
Reactions: AP514
like this
^ That is exactly why people should use reviews to determine relative performance between available coolers. It isn't rocket science to check a few review sites and go "Oh one of these coolers I'm interested in is 5 to 8C better than the other one I was looking at."
 
Ocellaris said:
^ That is exactly why people should use reviews to determine relative performance between available coolers. It isn't rocket science to check a few review sites and go "Oh one of these coolers I'm interested in is 5 to 8C better than the other one I was looking at."
Click to expand...
:confused:
I guess I failed. :(

My point is very few reviews tell us anything but how reviewer's system performs. Looking at them and saying "Cooler 'A' is 8-10c better than cooler 'B' is generally not even close to what the coolers differences are. It is only that specific system difference with cooler 'A' and cooler 'B'.

Case in point is the excellent Silver Arrow SB-E & IBE Extreme cooler as well as SST HE01.
In open air tests they are 6-10c better than other top coolers, including AIO cooler. But it case tests they are lucky to match other coolers.

Reason is extremely simple. They have fans moving twice as much air as other top coolers .. and in open air they all get air of similar temperature.

But inside of a case that is moving at best half as much air as they are, they are forced to recycle their own heated air .. which increases their intake air temp 6-16c higher than coolers with fans flowing half as much air. End result is CPU temps also going up dramatically making their "performance" look terrible.

The fact is it is not a lack of cooler performance, but a lack of case performance,
Case is only supply half as much air as cooler needs. Total case performance failure !!

But many readier only see the CPU temps and room temps. And have no idea what is actually going on inside of the case!

Edit: Check out "Ways to Better Cooling" 1st post is index, click on title to see it.
http://www.overclock.net/t/1491876/ways-to-better-cooling-airflow-cooler-fan-data/0_20
After you've read it, if you have any questions ask them here.
 
Last edited:
You don't need a scientific breakdown of cooler performance. If cooler A performs better than cooler B four times out of four on different sites with different test methods and CPUs, cooler A is tremendously likely to be better. Then it turns into questions about size, looks, and mounting method. You can get analysis paralysis if you try breaking down simple performance numbers too much.

No everyone needs to break down questions like "What if I live in a Vietnamese jungle" and "What if my case is only 3L and has close to zero airflow".
 
Last edited:
Ocellaris said:
You don't need a scientific breakdown of cooler performance. If cooler A performs better than cooler B four times out of four on different sites with different test methods and CPUs, cooler A is tremendously likely to be better. Then it turns into questions about size, looks, and mounting method. You can get analysis paralysis if you try breaking down simple performance numbers too much.

No everyone needs to break down questions like "What if I live in a Vietnamese jungle" and "What if my case is only 3L and has close to zero airflow".
Click to expand...
No, cooler A & B giving similar results in similar tests on 50 sites does not mean the testing criteria is correct. And if the testing is not done properly the results have little or no meaning.

I explained why it is not true and gave an example to show why.

But you obviously have no interest in truth or facts so feel free to believe what you want. Just don't play games trying to involve me in your game.


Another example is Frosty Tech. While the use good test procedure their maximum wattage testing in may of their tests is often not high enough to be applicable for oveclockers. For example their highest heat tests for Intel CPU when testing PH-TC14PE was only 150w .. when many intel CPUs are 125-135w stock . and 250-400w when overclocked. In their defence they now do 250w max load. But that is still now enough for serious oveclocke to use for cooler comparisons.
 
You must log in or register to reply here.
Back
Top