CPU Cooler Performance Tier List

[DarkReaper969]

3 minutes ago, CT007 said:

Where is the case fan performance list..?

Just keep an eye out for GamersNexus's videos, they recently got a massive fan tester unit, so there will be some reviews soon.

[Dogzilla07]

35 minutes ago, Tan3l6 said:

They are just made for different tasks, like CFM or dbA

Not really, they're all made for the same 4/5 tasks, and an industry standard P/Q curve exists to define it, which is seldom available. GamersNexus is about to change that.

 

A good static pressure fan is always better than an airflow fan, unless you can put your hand through the mounting slot where the fan will go (then u can use an airflow fan. Steve from GamersNexus calls this out 5 times every video about these things).

 

1. Filter+Mesh+Grill is the worst.

2. Filter+ seperate Mesh is as bad if not worse than Radiator if there's GamersNexus certified bad overlap between the two.

3. Heatsinks come with dense and not-dense fins and are are also significantly obstructive (it's a complete myth an "airflow" fan is good for them).

4. Radiators are radiators

5. integrated meshfilter is the least obstructive (if designed properly)

 

There's a gradation of Grill/Mesh as well.

 

1. Classic oldschool circle grill is the best

2. Hexagonal holes after at around ~72% letting air through

3. Pentagonal holes + irregular triangles and various rectangles

4. Square at about ~60% (like deepcool does)

5. Circles suck

 

of course the less material there is and more free space the better and changes the above a bit. it's been well document and available on Silverstone and Noctua websites for more than a decade now.

[Tan3l6]

7 minutes ago, Dogzilla07 said:

They're all made for the same 4 tasks, and an industry standard P/Q curve exists to define it, which is seldom available. GamersNexus is about to change that.

They are all made to move air, that is correct, only that for example throughput and silence and other factors are considered. Otherwise all the fans would be with same fan-blade design and overall efficiency.

[Dogzilla07]

16 minutes ago, Tan3l6 said:

They are all made to move air, that is correct, only that for example throughput and silence and other factors are considered. Otherwise all the fans would be with same fan-blade design and overall efficiency.

Not exactly true, re-read my post above. When a fan is designed, you can optimized it for any of my first 5 points to a degree. (to a point, more like a combination of the factors). CFM values are pointless if the testing doesn't take those factors into account. CFM without static pressure at a certain one of those 5 points is a misleading chart/number.

 

In general we don't use fans without obstructions anymore. So they all need to be designed with static pressure in mind (intake primarily of course). Additionally a slower moving fan with better static pressure will match an opposite fan, and inherently be less loud, so it's illogical to not make use of that.

 

You are correct that throughput and silence are the main factors and most important. But there is a plethora of nuance involved as to how you achieve that for the actual use-cases for the fans. It's not black and white.

 

https://www.silverstonetek.com/techtalk_cont.php?tid=wh12_008&area=usa

 

https://www.silverstonetek.com/techtalk_cont.php?tid=wh_chessis&area=usa

 

[Tan3l6]

@Dogzill

I'd, with some reservation would trust JayZ's video:

 

 

[Dogzilla07]

Did you just link JayZ, ... I'm done. Stop watching Jay for advanced stuff and especially fluid dynamic, air cooling and case cooling stuff (the amount of mistakes and inaccurate information he says is very high). it's GamersNexus or bust.

[Tan3l6]

2 minutes ago, Dogzilla07 said:

Did you just link JayZ, ... I'm done. Stop watching Jay for advanced stuff and especially fluid dynamic, air cooling and case cooling stuff. it's GamersNexus or bust.

I'm a simple man -  I see explanations I link the video.

he seems a bit spoon-feeding, yet it seems legit.

[Dogzilla07]

19 hours ago, Tan3l6 said:

I'm a simple man -  I see explanations I link the video.

he seems a bit spoon-feeding, yet it seems legit.

He's wrong ~30-55% of the time for that video (it's not all bad, he makes some good points, but even for those his explanations/reasonings are incorrect). And there's just plenty of plain wrong stuff (maybe I'm misremembering that it's 2/3rds, probably like 40% or so but it's not the video you wanna go for high-level stuff, I'll double-check and change the % later cause I don't wanna dis him too much, cause he's gotten a lot more diplomatic, accurate over the years, and open-minded). Jay knows the stuff he knows and had experience with, but he used to be very uncompromising when encountering  conflicting, new and changed stuff, he was very slow to adapt and change, and very defensive about his previous positions (especially when they were easily disproven).

 

Though thankfully since GamersNexus and HardwareUnboxed became big and started pushing the concepts of methodology and scientific methods, he and well all other youtubers have mellowed out a bit and gotten more accurate. in general you could say GamersNexus has put the fear of GOD (in the biblical sense) in every youtuber except Buildzoid (but that's only cause Buildzoid has been putting the fear of GOD into himself to be more accurate since he was born xD)

 

 I can't wait for GamersNexus to release videos, so I can just post their links, they've saved me 1,000s of hours of explaining and linking to productive discussions on forums with past and current reviewers, former and current industry professionals, and experienced users, overclockers and scientists.

 

And again like don't get me wrong, I have utmost complete understanding and sympathy for other youtubers, this is more work, and less money for them. Mostly incredibly more work. GamersNexus just ups the ante constantly. But for us consumers, GamersNexus is the best thing that happened since JonnyGuru and Aris.

[Tan3l6]

2 minutes ago, Dogzilla07 said:

He's wrong 70% of the time for that video.

OK, then i'm not a fan.
Seriously though - I appreciate your input!

[WoodenMarker]

On 3/2/2022 at 12:40 AM, arc.neiht said:

guys how about Jonsbo HX7280, what tier of this!?

Actually I wondering Jonsbo HX7280 vs Coolermaster ML240L V2 with i7 12700f for work(render corona, sketchup,...)

Added to the list. It seems similar to the Assassin III based on the few reviews available. 

https://coolenjoy.net/bbs/review/921910?p=5

https://www.chiphell.com/portal.php?mod=view&aid=26718&page=6

https://quasarzone.com/bbs/qc_qsz/views/1197474

Either would be fine but the HX7280 seems better than the ML240L V2. 

15 hours ago, CT007 said:

Where is the case fan performance list..?

Performance is roughly higher rpm > lower rpm. 

When accounting for noise, what's considered quiet becomes subjective. Noise can be measured in decibels but different frequencies can be more or less noticeable to different people. Acoustics can also vary depending on where a fan is mounted. 

[Dogzilla07]

All in all, like the hunch i had , I was wrong, miss-remembered, and overreacted. it's not nearly that bad, He makes many very good good points, gives much good information, says some exact stuff i did (so actually proves my points few posts back), and it's at most 15-25% wrong, but probably even less than that.

My "it's not good for high-level stuff" comment hasn't aged well, and i'm gonna eat my hat on that one. It's plenty good for high-level stuff.

 

Most of the wrong stuff is minor, Only the 200x25mm fan part of the video is a big wrong. And the bearing explanation is well, i've listened to it many times, and in essence it's all correct, just sometimes a mixed name or two for a bearing type made it slightly confusing. But reading what i wrote below, i'm confused as well, and I think i made some mistakes with rifle bearings. Either way it's a complex situation with those fan bearing names. I was posting wrong stuff about bearings for years in the past, it isn't until recently (~2-3 years) that my understanding of them has gotten good enough.

 

starting from 2:10 especially 2:30 He accurately explains the use-cases i mentioned as well, just not all-of them

2:54 he mentioned the gap again correct (but pitch, and specific angles as well also have great impact)

3:03 he mentiones air is a fluid (like i did)

4:00 is a partial mistake. the blow-back he's explaining also (and mostly) happens outside the frame (a loop/vortex happens), and the problem is the thin frame like on those corsairs (which is one of the reasons why people recommend taping all holes around intake fans. he's partially right here.

6:15 the mistake i remembered. he says the smaller the fan the greater the static pressure, and that is correct in general without taking noise into question, but depending on the (depth, fatness of the frame), the type of material used, when you compare the fans noise-normalized a 140mm can have greater static pressure than a 120mm.

6:34 mistake again. What he says is true now because of Liquid-Crystal Polymer 120mm fans, but wasn't true before. There are 140mm fans that give the same static pressure for the same noise (he says lower RPM, but this is where the mistake happens you cannot 1:1 compare RPMS between 120mm and 140mm,even though they spin slower they are competitive or beat faster spinning 120mm, which he accurately explains afterwards). RPM-based testing is very flawed, as GamersNexus mentions all the time).

6:45 correct, but he assumes motherboards couldn't handle 140mm fans, but that's bad 140mm fans, enthusiasts were using niche 140mm fans that didn't exceed fan headers, and there were plenty 120mm fans that did "exceed" fan headers.

7:36 swept comment, awesome, he made very good points

8:20 vardar comment, proven by VSG on thermalbench testing (he does fan testing on tomshardware, Vardar + Eloops= top performance), VSG as a reviewer very high quality methodology, GN before GN. But vardar have motor noise problems, hence new ones incoming.

9:10 all great info

10:00 Innacuracies here. Mentioned nice slow cool volume of air = cool enviroment, which just isn't true. Top of the line 140mm fans beat those thin 200x25mm fans any day, LCP 120mm fans beat them as well. When noise-normalized and against obstructions those 200x25mm fans are just anemic. It's been discussed and proven ad infinutum over the past 7 years or so. They need to be fatter.

10:55 200x25mm do not give the best exchange of air to noise to volume. GamersNexus is about to bust this myth wide open before summer probabaly (only the Noctua 200mm is ok-ish, but still inferior to top 140mm and top LCP 120mm fans.

12:05 partially correct, but the header burning problem, does not come from how much the fan pulls when running, but how much the fan pulls when starting up (that's the danger), and often those numbers are missing on official websites (as fans start up with higher amparage than they end up when running). All good info about NZXTs

12:54 open airflow infront comment, like i mentioned in this thread.

13:27 not all fans move in and out, some high quality fans are extremely stabilized like (Scythe GT, and others). After than all good info

15:05 You can, but don't want to just put mineral oil, you wanna put grease, specifically silicone or graphite grease (graphite is abrasive, so silicone grease is better). and then because you're putting grease, it's fine to do a drop or two of sewing machine oil.

15:19 Not all fluid-bearings are made the same, the bearing they're based, basic sleeve, yes that fails the fastests, but Be Quiet! and Noctua and Thermalright and Deepcool and some Cooler master, and many others have some really cool advanced self-lubricating bearings

15:40 good info for ball-bearing (but maglev i think is more expensive). Big problem with ball bearings is that they produce a specific noise most people can't handle

16:03 sleeve is not usually a hybrid but i think it can be (but not in PC space), fluid-bearing is just an advanced sleeve bearing. 

16:40 rifle bearing is an advanced sleeve, correct unlike above, and some "hydro" mentioned bearings are advanced rifle bearings actually

17:10 sleeve + rifle + hydro + loop + fluid = not for different directions (advanced loop fluid and hydro are ok), but basic ones are not good when mounting horizontal

17:10 ball bearings are, ball-bearings are specifically designed for horizonal (PSU use, GPU use). 

horizontally bad = fluid = hydro = loop = sleeve = rifle

horizontally good = ball, magnetic

however top of the line hydro and fluid and loop can do horizontally no problem (and so do most rifle bearings, because rifle=hydro depending on the manufacturer)

17:57 he correctly mentions ball bearings can do different orientation better. good comment for rifle

18:40 big no for sleeve ( but I think he is mixing up the nomenclature a bit and in essence he's explaining correctly, but labeled some bearings wrong)

20:26 high static pressure comment

 

[NighhhT]

ID-Cooling SE-225-XT and ID-Cooling SE-226-XT are missing from the list:

 

https://www.tweaktown.com/reviews/9867/id-cooling-se-225-xt-black-cpu-cooler/index.html

https://www.tweaktown.com/reviews/10022/id-cooling-se-226-xt-black-cpu-cooler/index.html

 

The SE-226-XT seems to have the same performance as the upcoming Scythe Mugen5 Rev.C (Japanese, no subs):

 

[viceice]

Hello, asking for a friend here, my friend has a H45 cooler with only one fan (120 mm) and wants to build a system with a 5600x

 

seeing it as Tier 4, one would think its a decent cooler for 5600x but our researched showed that even stock fan that comes with 5600x would be better for it

 

what do you think? 

[WoodenMarker]

3 hours ago, viceice said:

Hello, asking for a friend here, my friend has a H45 cooler with only one fan (120 mm) and wants to build a system with a 5600x

seeing it as Tier 4, one would think its a decent cooler for 5600x but our researched showed that even stock fan that comes with 5600x would be better for it

what do you think? 

The Wraith Stealth included with the 5600x is currently in Tier 10. It performs worse than most aftermarket solutions and much worse than the H45.

I haven't seen any benchmarks of the H45 on Ryzen since that was released years later than the cooler but the most similar analogue would be the 6950x. 

https://www.clubedohardware.com.br/analises/refrigeracao/teste-do-cooler-corsair-h45-r36640/?nbcpage=6

https://www.clubedohardware.com.br/forums/topic/1178980-teste-water-coolers-vs-air-coolers-resultado/

https://thermalbench.com/2016/09/27/corsair-hydro-series-h45-aio-cpu-cooler/5/

 

There are some amazon reviews of the cooler on other Ryzen cpus

https://www.amazon.com/Corsair-Hydro-Processor-Liquid-Cooler/product-reviews/B01FSMWFQG/ref=cm_cr_arp_d_viewopt_kywd?ie=UTF8&reviewerType=all_reviews&pageNumber=1&filterByKeyword=h45+ryzen

 

What tests or reviews are you looking at that indicate otherwise?

[viceice]

6 hours ago, WoodenMarker said:

The Wraith Stealth included with the 5600x is currently in Tier 10. It performs worse than most aftermarket solutions and much worse than the H45.

I haven't seen any benchmarks of the H45 on Ryzen since that was released years later than the cooler but the most similar analogue would be the 6950x. 

https://www.clubedohardware.com.br/analises/refrigeracao/teste-do-cooler-corsair-h45-r36640/?nbcpage=6

https://www.clubedohardware.com.br/forums/topic/1178980-teste-water-coolers-vs-air-coolers-resultado/

https://thermalbench.com/2016/09/27/corsair-hydro-series-h45-aio-cpu-cooler/5/

 

There are some amazon reviews of the cooler on other Ryzen cpus

https://www.amazon.com/Corsair-Hydro-Processor-Liquid-Cooler/product-reviews/B01FSMWFQG/ref=cm_cr_arp_d_viewopt_kywd?ie=UTF8&reviewerType=all_reviews&pageNumber=1&filterByKeyword=h45+ryzen

 

What tests or reviews are you looking at that indicate otherwise?

sadly, we couldn't find any actual benchmarks.

 

"I would say not, I would highly recommend you to use the stock cooler instead of wasting money on that thing."

 

 

 

"120mm liquid coolers suck."

 

80c on r5 3600 seems a lot, friend wants temps near 60-70. we checked stock cooler benchmarks for the 5600x and in heavy games it seems to hover around 80c, which makes me question the h45's validity.  h45 is also much much cheaper than the coolers within that tier. after all, there are lots of coolers in tier 4 that are suggested reliably for 5600x. by that logic, h45 should manage the 5600x fine, but I have doubts

 

if its really going perform similar to the, say, arctic espots 34 duo (tier 6), then i will advise my friend to keep his h45. his budget allows tier 5-6 CPUs at the moment and he already has a h45. if h45 is indeed similar or a bit of better compared to those coolers, then he should keep using it, no?

 

 

[WoodenMarker]

8 hours ago, viceice said:

sadly, we couldn't find any actual benchmarks.

"I would say not, I would highly recommend you to use the stock cooler instead of wasting money on that thing."

 

"120mm liquid coolers suck."

 

80c on r5 3600 seems a lot, friend wants temps near 60-70. we checked stock cooler benchmarks for the 5600x and in heavy games it seems to hover around 80c, which makes me question the h45's validity.  h45 is also much much cheaper than the coolers within that tier. after all, there are lots of coolers in tier 4 that are suggested reliably for 5600x. by that logic, h45 should manage the 5600x fine, but I have doubts

 

if its really going perform similar to the, say, arctic espots 34 duo (tier 6), then i will advise my friend to keep his h45. his budget allows tier 5-6 CPUs at the moment and he already has a h45. if h45 is indeed similar or a bit of better compared to those coolers, then he should keep using it, no?

If you already have the H45, there's no reason to downgrade.

 

The thread that mentioned 80c wasn't with the H45. Seems to be the stock cooler.

People hate on 120mm AIOs for some reason despite how they perform in benchmarks.

[Dogzilla07]

16 hours ago, WoodenMarker said:

People hate on 120mm AIOs for some reason despite how they perform in benchmarks.

It's very easy to get horrible noise-normalized performance for even 12100 (let alone 12400F/5600x) on 120/140mm AIOs with bad fans, that's why (and a smaller sub-set have inadequate pumps, but quite smaller). How good the fan is comes to light/into play much faster than on bigger radiators.

 

Second reason is that without soft-modding and major tinkering 120/140mm radiators are inferior to top-down blower coolers in many Small Form Factor cases (especially when air is combined with the new Noctua NA-FD1 fan duct). As has popped up in dozens upon dozens of Machines&More, OptimumTech, HardwareCanucks videos, and hundreds upon hundreds of reddit posts.

[NighhhT]

On 3/5/2022 at 4:35 PM, NighhhT said:

ID-Cooling SE-225-XT and ID-Cooling SE-226-XT are missing from the list:

 

https://www.tweaktown.com/reviews/9867/id-cooling-se-225-xt-black-cpu-cooler/index.html

https://www.tweaktown.com/reviews/10022/id-cooling-se-226-xt-black-cpu-cooler/index.html

 

The SE-226-XT seems to have the same performance as the upcoming Scythe Mugen5 Rev.C (Japanese, no subs):

 

No point on putting the SE-225-XT and SE-226-XT in the tier list? @WoodenMarker

[WoodenMarker]

4 hours ago, NighhhT said:

No point on putting the SE-225-XT and SE-226-XT in the tier list? @WoodenMarker

Would love to but need more data / reviews.

[Jeppes]

On 3/10/2022 at 11:01 PM, NighhhT said:

No point on putting the SE-225-XT and SE-226-XT in the tier list?

Those reviews are pretty much garbage as done with a 3600x. At the minimum you would love to see a modern 8-core with 140-150W:s, better if closer to 200W to get larger gaps between best and mediocre coolers.

[DarkReaper969]

Why is the Dark Rock Pro 4 in Tier 4? Isn' t it one of the best performing coolers?

[WoodenMarker]

8 hours ago, DarkReaper969 said:

Why is the Dark Rock Pro 4 in Tier 4? Isn' t it one of the best performing coolers?

It's quiet but far from the best performer. It's an average performer among the 120-140mm dual tower air cooler competition. 

[DarkReaper969]

10 hours ago, WoodenMarker said:

It's quiet but far from the best performer. It's an average performer among the 120-140mm dual tower air cooler competition. 

Ah alright thanks, than maybe some people just praised the performance of it a bit too much. 

[Alex13]

Any clue how long it should take before the Fuma 2 rev. B is added to the list? No reviews available yet and I'm wondering if the new fans with 1500 max RPM will make a big enough difference to move the cooler to tier 3 and close the gap with NH-D15. Also hope the noise normalized performance profile will be as good or better across the whole RPM range.

[WoodenMarker]

6 hours ago, Alex13 said:

Any clue how long it should take before the Fuma 2 rev. B is added to the list? No reviews available yet and I'm wondering if the new fans with 1500 max RPM will make a big enough difference to move the cooler to tier 3 and close the gap with NH-D15. Also hope the noise normalized performance profile will be as good or better across the whole RPM range.

No idea how long it would be. The NH-D15S is already better than the Fuma 2 so the higher rpm Fuma 2 Rev. B might match the NH-D15S.