CPU Cooler Performance Tier List

[WoodenMarker]

On 10/20/2019 at 9:42 AM, ThiccSmough said:

It seems that there are few reviews about Thermalright True Spirit 140 Power.

On the other forum, some people strongly recommended TS140P due to its low price while being competitive compared to D14 or D15.

I'm assuming Overclock.net?

The TS140P is a great cooler. On paper, it should outperform coolers like the Macho Rev. B and I expected to put it higher on the list but the results don't show that consistently between other people's tests. I've read that it particularly benefits from lapping. The convex base can be great at times but might not be the best fit for most cpus which might be what's hurting its consistency. 

The out-of-box performance isn't as good as the NH-D14 or NH-D15. It can get close but it loses more often than not. It's better compared to the similar NH-U14S which will also get close to the NH-D14 or NH-D15 sometimes. 

[Juular]

Can you clear out a few things ?

- Why Scythe Ninja 5, Mugen 5 and Thermalright Macho X2 with two fans are in the same tier as their single-fan versions ? Shouldn't they be moved one tier up ?

- As far i'm aware be quiet! Dark Rock Pro 4 are pretty much on par with Noctua NH-D15 in all aspects so they should be in the same tier ?

 

Also your tier naming convention is very bizarre, 'Kings of Whoosh \ Sploosh' ? What does that even mean ? And well, tier 3 aren't all dual-towers nor tier 4 140mm. Maybe you should put some TDP range in there or CPUs they're recommended for instead ? Like tier 1 - '200-250W+ HEDT CPUs \ extreme OC', tier 2 - '150-200W 12+ core high OC \ 6\8-core prolonged high-load & low-noise operation', tier 3 - '6\8-core mild OC \ 4-core high OC \ low noise AAA gaming \ high-load workstation', tier 4 - '4-core mild OC \ AAA gaming', tier 5 - 'non-demanding games \ low-load workstation', tier 6 - 'better stock cooler replacement \ low noise office PC', tier 7 - 'stock cooler replacement \ office PC', tier 8 - 'low profile \ only when size is constraint', tier 9 - 'small downdrafts \ for HTPC or if it's free'.

[WoodenMarker]

4 hours ago, Juular said:

Can you clear out a few things ?

- Why Scythe Ninja 5, Mugen 5 and Thermalright Macho X2 with two fans are in the same tier as their single-fan versions ? Shouldn't they be moved one tier up ?

- As far i'm aware be quiet! Dark Rock Pro 4 are pretty much on par with Noctua NH-D15 in all aspects so they should be in the same tier ?

 

Also your tier naming convention is very bizarre, 'Kings of Whoosh \ Sploosh' ? What does that even mean ? And well, tier 3 aren't all dual-towers nor tier 4 140mm. Maybe you should put some TDP range in there or CPUs they're recommended for instead ? Like tier 1 - '200-250W+ HEDT CPUs \ extreme OC', tier 2 - '150-200W 12+ core high OC \ 6\8-core prolonged high-load & low-noise operation', tier 3 - '6\8-core mild OC \ 4-core high OC \ low noise AAA gaming \ high-load workstation', tier 4 - '4-core mild OC \ AAA gaming', tier 5 - 'non-demanding games \ low-load workstation', tier 6 - 'better stock cooler replacement \ low noise office PC', tier 7 - 'stock cooler replacement \ office PC', tier 8 - 'low profile \ only when size is constraint', tier 9 - 'small downdrafts \ for HTPC or if it's free'.

There are definitely coolers that are better than others in each tier. If each tier just meant coolers that perform the same, there would be a tier for pretty much every cooler. 

The tiers are just arbitrary groups that can separate the coolers as much as possible given the data. With more data / confidence, there could be a few more separations / tiers. 

 

The Ninja 5 just has two fans. There are older revisions that use fewer fans but it's not the same heatsink with a different number of fans. The Mugen 5 is similar. It only comes with one fan. The closest to a 2 fan version is the Mugen 5 PCGH which comes with lower rpm fans. 

 

The DRP4 is far from being on par with the NH-D15 for performance in that there are plenty of tests showing the NH-D15 performing better than not. The DRP4 is already outperformed by the NH-D15S which has one single fan..

This video has some good comparisons over the differences between some large coolers: 

 

As mentioned in the intro, "The tiers and titles are arbitrary and mainly to make it easier to sort by performance." 

There's a trade off between having consistent and useful titles. 

 

Regarding your suggestions:

HEDT is kind of a platform and I don't want to confuse that with an arbitrary section title.

Noise isn't considered for the tier list. The list concerns performance only and adding noise makes it both more complicated than it already is, adds some subjectivity, and makes it less accurate without a large influx of extra data. 

The quality of the cooler outside of what's reflected in the performance itself also isn't considered for the same reasons and prevents any sort of recommendations being made in the same list. 

Recommendations for any particular cpu type as categories are necessarily inaccurate since different coolers perform differently in different systems for various reasons. At the end of the day, the point of the tier is to explain that a cooler in it is usually better than those in lower tiers. There are no promises to what is enough to cool a cpu -- please refer to Gamers Nexus' video below that elaborates on this problem which can be summarized by the existence of TDP. 

 

As mentioned in comments on the first page, TDP is not a great measurement especially for accuracy. 

Gamers Nexus recently uploaded a good video going over this: 

 

[Juular]

2 hours ago, WoodenMarker said:

As mentioned in comments on the first page, TDP is not a great measurement especially for accuracy.

I'm talking about TDP exactly in the way Tech Jesus does, in a strict sense, electrical power = thermal dissipation power. If CPU consumes 100W under certain load then it dissipated almost exact same thermal power. I agree that it's probably not a good way to chose cooler as neither CPU manufactuers nor cooler manufacturers define TDP in that way. So maybe just some broad CPU recommendations disregarding noise performance if you don't want to make it more complicated, as your current titles are neither consistent nor useful

Quote

The Mugen 5 is similar. It only comes with one fan.

Yeah, i've missed that.

As of DRP4, just first 3 reviews i found :

Tweaktown - 2°C & 2 dBa (both worse) difference both in PWM & 12V mode, well within measurement error.

Guru3D - 7 dBa lower noise with 3° worse thermals, again seems to be on par if we normalize it by noise.

TechPowerUp - 1 dBa higher noise on 25% & 50% RPM and 3 dBa lower noise on 100% RPM. Thermals are either exact same or slightly lower\higher (within 2-3°C difference). It's not clear at which RPMs tests were performed but either way it seems to be pretty much on par on average.

Respectfully, i don't consider Youtube reviews to be worthy, if there's not enough information for a full blown article as GamersNexus do. Most of them are half-assed at best, often straight up misleading.

As far of quality, be quiet! stuff is very good, i don't have their coolers but i have 3 of their SW3 fans (same as on their coolers) and they're fantastic, so i still don't see why they should be one tier lower than Noctua stuff.

[Algebruh]

CPU coolers' TDP doesn't really mean anything without temperature delta.

 

Let's say A and B cooler both are capable of dissipating 200W heat load.

 

The temperature delta of A cooler is 40 ℃ while the temperature delta of B cooler is 50 ℃.

 

It is clear that A cooler is better than B cooler, even though both coolers are capable of dissipating 200W heat load.

 

[WoodenMarker]

13 hours ago, Juular said:

I'm talking about TDP exactly in the way Tech Jesus does, in a strict sense, electrical power = thermal dissipation power. If CPU consumes 100W under certain load then it dissipated almost exact same thermal power. I agree that it's probably not a good way to chose cooler as neither CPU manufactuers nor cooler manufacturers define TDP in that way. So maybe just some broad CPU recommendations disregarding noise performance if you don't want to make it more complicated, as your current titles are neither consistent nor useful

Yeah, i've missed that.

As of DRP4, just first 3 reviews i found :

Tweaktown - 2°C & 2 dBa (both worse) difference both in PWM & 12V mode, well within measurement error.

Guru3D - 7 dBa lower noise with 3° worse thermals, again seems to be on par if we normalize it by noise.

TechPowerUp - 1 dBa higher noise on 25% & 50% RPM and 3 dBa lower noise on 100% RPM. Thermals are either exact same or slightly lower\higher (within 2-3°C difference). It's not clear at which RPMs tests were performed but either way it seems to be pretty much on par on average.

Respectfully, i don't consider Youtube reviews to be worthy, if there's not enough information for a full blown article as GamersNexus do. Most of them are half-assed at best, often straight up misleading.

As far of quality, be quiet! stuff is very good, i don't have their coolers but i have 3 of their SW3 fans (same as on their coolers) and they're fantastic, so i still don't see why they should be one tier lower than Noctua stuff.

TDP can't be derived from review results. What you're suggesting is a 'universal standard' which doesn't really work. 

 

The titles are summaries / generalizations. In the real world, cooler performance can only be generalized since coolers have to work with different situations. 

The only information conveyed by the list is that coolers in a tier tend to perform better than coolers in lower tiers. Hence the title.

In that same sense, the tier titles summarize High end 140mm's tend to perform better than other 140mm's and high end 120mm's, etc. 

 

Please show where the margin of error is stated by Tweaktown or any other reviews. The margin of error is calculable and you don't get to just pick what it is. 

If I'm remembering correctly, the margin of error if stated at all by Tweaktown was 1c.

A result within margin of error doesn't mean that it's a non result either. Margin of error is usually relevant only when the results don't match with most others.

The list compares raw performance meaning max rpm and as heavy loads as possible. 

Please link articles when referring to them. 

 

Feel free to look at other sources if you don't like Youtube reviews. I'll continue to use them as long as the data is useful. 

The DRP4 is lower because it performs worse. If you can find sources that show that this is indeed not the case and are more compelling than results that show otherwise, then it might be moved up or the NH-D15 moved down. 

[Juular]

58 minutes ago, WoodenMarker said:

If you can find sources that show that this is indeed not the case and are more compelling than results that show otherwise, then it might be moved up or the NH-D15 moved down. 

Are these three reviews i mentioned not enough ? I wasn't clear enough maybe, not margin of error per se but rather measurement variables, you will not find two reviews that show exact same values and across these three reviews i mentioned these values change in favor to both sides so you can pretty much say that on average neither of them have a significant advantage. And that's just first three reviews i found on google, there might be more of them. But if you still find this data not enough that's fine, this list is your opinion and i'm not gonna argue to the death about it.

5 hours ago, ThiccSmough said:

CPU coolers' TDP doesn't really mean anything without temperature delta.

I wasn't talking about coolers TDP but CPU TDP, two CPUs with same thermal dissipation power under same type of load can be pretty much generalized by that number regardless of manufacturer, amount of cores, frequency etc. Unless they have large difference in IHS area like TR4\LGA2011 vs AM4\LGA115x. In that regard you can probably need a separate tier for HEDT CPUs with giant IHS with coolers that can cover it. And back to the noise in the tier name, i wasn't talking about some exact numbers but rather like say, Noctua NH-D15 for gaming is way overkill you're not gonna see more than 75W load in gaming regardless of CPU but it'll be rather very quiet at that, quieter than lower tier coolers so they still can be recommended for gaming if you want very quiet system, otherwise smth from one or two tiers lower will be just fine.

[WoodenMarker]

28 minutes ago, Juular said:

Are these three reviews i mentioned not enough ? I wasn't clear enough maybe, not margin of error per se but rather measurement variables, you will not find two reviews that show exact same values and across these three reviews i mentioned these values change in favor to both sides so you can pretty much say that on average neither of them have a significant advantage. And that's just first three reviews i found on google, there might be more of them. But if you still find this data not enough that's fine, this list is your opinion and i'm not gonna argue to the death about it.

 

I wasn't talking about coolers TDP but CPU TDP, two CPUs with same thermal dissipation power under same type of load can be pretty much generalized by that number regardless of manufacturer, amount of cores, frequency etc. Unless they have large difference in IHS area like TR4\LGA2011 vs AM4\LGA115x. In that regard you can probably need a separate tier for HEDT CPUs with giant IHS with coolers that can cover it. And back to the noise in the tier name, i wasn't talking about some exact numbers but rather like say, Noctua NH-D15 for gaming is way overkill you're not gonna see more than 75W load in gaming regardless of CPU but it'll be rather very quiet at that, quieter than lower tier coolers so they still can be recommended for gaming if you want very quiet system, otherwise smth from one or two tiers lower will be just fine.

What are you referring to by values? What I'm looking for is only the answer of which cooler performs better. It doesn't matter how large or small the difference is. If a cooler is better, it's in the same tier or higher on the list. 

If you think it's a matter of opinion and you disagree, please let me know where you think a cooler that consistently performs better is placed lower than a cooler that performs worse. 

 

Cooler and cpu tdp are the same and serve the same purpose. 

You say the NH-D15 is overkill for gaming. Plenty of people disagree. That is an opinion and not a true summary / generalization of a tier's content. 

[Juular]

9 hours ago, WoodenMarker said:

What are you referring to by values? What I'm looking for is only the answer of which cooler performs better. It doesn't matter how large or small the difference is. If a cooler is better, it's in the same tier or higher on the list.

Damn, for the third time. These three review sites i mentioned have data on both these models in question.

https://www.tweaktown.com/reviews/8842/quiet-dark-rock-pro-4-cpu-cooler-review/index6.html

https://www.guru3d.com/articles_pages/be_quiet_dark_rock_pro_4_review,7.html

https://www.techpowerup.com/review/be-quiet-dark-rock-pro-4/6.html

9 hours ago, WoodenMarker said:

Cooler and cpu tdp are the same and serve the same purpose.

There's no simple way to measure cooler's TDP, every manufacturer can come out with random formula to calculate it with different values so you can't compare it. On the other hand you can define CPU's TDP as simple as being equal to it's maximum power draw which will be almost equal to it's dissipated heat.

As of tiers, you can define say 150W tier as where all coolers meeting it don't exceed some dT (say 60° C over ambient) cooling a CPU that draws and therefore dissipates 150W. But without massive amount of data that's not quite possible, so again, i acknowledge that TDP is of no use when defining tiers unless you can test these coolers yourself.

[WoodenMarker]

9 hours ago, Juular said:

Damn, for the third time. These three review sites i mentioned have data on both these models in question.

https://www.tweaktown.com/reviews/8842/quiet-dark-rock-pro-4-cpu-cooler-review/index6.html

https://www.guru3d.com/articles_pages/be_quiet_dark_rock_pro_4_review,7.html

https://www.techpowerup.com/review/be-quiet-dark-rock-pro-4/6.html

There's no simple way to measure cooler's TDP, every manufacturer can come out with random formula to calculate it with different values so you can't compare it. On the other hand you can define CPU's TDP as simple as being equal to it's maximum power draw which will be almost equal to it's dissipated heat.

As of tiers, you can define say 150W tier as where all coolers meeting it don't exceed some dT (say 60° C over ambient) cooling a CPU that draws and therefore dissipates 150W. But without massive amount of data that's not quite possible, so again, i acknowledge that TDP is of no use when defining tiers unless you can test these coolers yourself.

Sure. there's data. If by values never being the same, you mean the temps, that's fine. They will never be the same and it doesn't matter as long as the data can be used to determine which cooler performs better. 

 

Even when testing coolers myself, the tests done wouldn't be extensive enough to cover the variety in real life use. That's why I use real life results in the form of many tests and reviews as opposed to a few synthetic, 'clean' results. 

For coolers as popular as the DRP4 and NH-D15, the results aren't from a measly 3 direct comparisons but closer to up to 100 or more direct and indirect comparisons. 

[Juular]

3 hours ago, WoodenMarker said:

Sure. there's data. If by values never being the same, you mean the temps, that's fine. They will never be the same and it doesn't matter as long as the data can be used to determine which cooler performs better. 

 

Even when testing coolers myself, the tests done wouldn't be extensive enough to cover the variety in real life use. That's why I use real life results in the form of many tests and reviews as opposed to a few synthetic, 'clean' results. 

For coolers as popular as the DRP4 and NH-D15, the results aren't from a measly 3 direct comparisons but closer to up to 100 or more direct and indirect comparisons. 

I still don't see your point, you've asked for sources, i provided. Are you saying that you have close to 100 of other sources or what ?

[WoodenMarker]

On 10/25/2019 at 11:28 PM, Juular said:

I still don't see your point, you've asked for sources, i provided. Are you saying that you have close to 100 of other sources or what ?

I'm referencing 100+ reviews for the placement of the DRP4 and NH-D15, yes. 

Noctua currently links 300+ reviews for the NH-D15 on their website and there are more online that's not listed: https://noctua.at/en/nh-d15/review

Plenty of appearances for the NH-D15 are from reviews of other coolers. 

All of the data from those sites has been and is considered if usable / relevant. 

 

I wouldn't say I have those sources since I haven't archived them and I might not ever. Many of the older sites have been down since I've accessed them as well. 

I began manually cataloging review links and results a while back but have stopped due to the amount of time it takes. I also couldn't get my head around the math needed for accurate rank aggregation when working with some friends on making a program to rank the coolers. We got something basic working but the math wasn't adequate.

Here's a small snippet of what came from a few reviews from 50+ sources and 300+ individual reviews / tests. The index being incomplete yet having 1300+ coolers / configurations is what made the manual data entry particularly daunting.

Spoiler

 

There are plenty of reviews sites listed in the Resources section of the list. 

[RandomDavid]

Hey, I'm new here. Is Gammaxx Gte RGB good or Cryorig H7 or cooler master hyper 212 black edition? Any recommendation for air cooler??

[WoodenMarker]

4 minutes ago, RandomDavid said:

Hey, I'm new here. Is Gammaxx Gte RGB good or Cryorig H7 or cooler master hyper 212 black edition? Any recommendation for air cooler??

They're all fine and fairly similar. 

What are your specs and why are you looking at these coolers in particular?

Are there any features you're looking for in a cooler?

Where are you shopping / located? Budget?

[RandomDavid]

I'm currently earning money to build a new pc so I'm trying to find a best budget cpu cooler.

CPU: R5 3600

GPU: PowerColor Rx 5700xt Red devil

Motherboard: Msi Meg x570 unify

RAM: Corsair vengeance Rgb Pro 16gb (2x8) 3200 mhz

PSU: Corsair Txm 650w

For the cooler feature... not that much actually just want them to sync rgb only and not too hot.

I'm from malaysia where I use Lazada to check prize and went Pc shop (known as Lowyat).to buy.

I don't want to overspend to much on cpu cooler since i have put most of my money on gpu :'(.

[WoodenMarker]

On 11/2/2019 at 12:06 AM, RandomDavid said:

For the cooler feature... not that much actually just want them to sync rgb only and not too hot.

I don't want to overspend to much on cpu cooler since i have put most of my money on gpu :'(.

Have you considered using the stock cooler first? 

It's not amazing but is adequate for keeping temps in check. You can always decide to replace the cooler later if you want lower temps or less noise. 

If you don't like the look of it, that's not a bad reason to get a different cooler to begin with. What case are you using?

[RandomDavid]

17 hours ago, WoodenMarker said:

Have you considered using the stock cooler first? 

It's not amazing but is adequate for keeping temps in check. You can always decide to replace the cooler later if you want lower temps or less noise. 

If you don't like the look of it, that's not a bad reason to get a different cooler to begin with. What case are you using?

I did consider about it first but not so sure about whether it can handle the CPU. I'm using a Gaming Freak 50s-Neo ( Malaysia), although I might switch to lian li lancool 2 ( haven't come out yet). I will check the temps before install in.

[FK2tame]

Hi, is there Dual Tower air cooler that didn't exceed 160mm height?

i found Noctua NH-D14 but i wanted to have white color scheme and sadly chromax only support NH-D15 rite?

 

[WoodenMarker]

On 11/5/2019 at 7:06 AM, RandomDavid said:

I did consider about it first but not so sure about whether it can handle the CPU. I'm using a Gaming Freak 50s-Neo ( Malaysia), although I might switch to lian li lancool 2 ( haven't come out yet). I will check the temps before install in.

It should be fine for keeping temps in check. As mentioned, you can always decide to replace the cooler later if you want lower temps or less noise. 

6 hours ago, FK2tame said:

Hi, is there Dual Tower air cooler that didn't exceed 160mm height?
i found Noctua NH-D14 but i wanted to have white color scheme and sadly chromax only support NH-D15 rite?

There are plenty.

The NH-D15 Chromax heatsink is 160mm tall. The listed 165mm spec is with the front fan raised for 32mm of clearance for the ram. With the middle fan lowered and front fan removed or moved to the rear, the cooler can potentially be 160mm tall. That being said, it's taller if you plan to use the Chromax heatsink covers. 

 

What cpu, case, and ram are you using and what kind of features are you looking for in a cooler?

Where are you shopping / located? Budget?

[RandomDavid]

1 hour ago, WoodenMarker said:

It should be fine for keeping temps in check. As mentioned, you can always decide to replace the cooler later if you want lower temps or less noise.

Okay, thanks for replying

[FK2tame]

2 hours ago, WoodenMarker said:

What cpu, case, and ram are you using and what kind of features are you looking for in a cooler?

Where are you shopping / located? Budget?

I'm using i5-8600k, case darkFlash DLM21 Mesh white color, and currently i have 2 sticks of ram without heatsink, 2 sticks of Vengeance LED.

I'm located at Malaysia, and budget below $80. Actually i'm trying to imitate this built. Something closer enough should do. (in this video, he's using Noctua NH-D15 along with Chromax, inside Fractal Design Meshify C)

[WoodenMarker]

21 hours ago, FK2tame said:

I'm using i5-8600k, case darkFlash DLM21 Mesh white color, and currently i have 2 sticks of ram without heatsink, 2 sticks of Vengeance LED.
I'm located at Malaysia, and budget below $80. Actually i'm trying to imitate this built. Something closer enough should do. (in this video, he's using Noctua NH-D15 along with Chromax, inside Fractal Design Meshify C)

The minimalist flat look in white is rather unique and there are already only a few options in white.

Neptwin white is good: https://www.lelong.com.my/deepcool-neptwin-white-air-cpu-cooler-easyit2u-215641026-2020-11-Sale-P.htm

If you don't need much cooling performance but just want white, alternatives include

Hyper 212 LED Turbo WE: https://www.lelong.com.my/cooler-master-hyper-212-led-turbo-white-edition-cpu-cooler-netstorecommy-I6253355-2007-01-Sale-I.htm

SE-214L: https://www.lelong.com.my/id-cooling-se-214l-snow-cpu-cooler-130mm-pwm-fan-lingloong-I5705420B-2007-01-Sale-I.htm

 

If you're not satisfied with any of these options, it's not very difficult to get a different cooler and mask / paint the top fin white. 

[balivyta]

dont really get how to use this chart...

which one is better, dark rock 4, fuma 2 or neptwin?

[WoodenMarker]

On 11/8/2019 at 9:36 AM, balivyta said:

dont really get how to use this chart...

which one is better, dark rock 4, fuma 2 or neptwin?

Higher on the list means higher performance. There are other factors that may make a cooler better depending on what features you want that aren't represented in the list such as acoustics or compatibility. 

So out of the 3 the best performing is the Fuma 2 followed by the Neptwin and DR4. 

[DarkZentrax]

Hello guys! I ask you a question, I found almost no information about the AeroCool Pulse L240F RGB, what do you think? it's good? For OC 8700K.