• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Have AIOs killed custom loops?

LOL ok.. so you believe that all comparative performance testing with any type of hardware should also include that the person installing said hardware might have installed one of the items of hardware incorrectly?
Don't be pedantic. We're not talking about standard computer hardware. We're talking about mostly specialized water based fluid circulation hardware. They are in very different classes and types of equipment. And yes, the competency of the installer is relevant in all cases.

AIO's are greatly easier to install and require MUCH less experience, thus much lower chance of screwing it up.

This is very clear basic logic, and really just common sense, so I'm surprised you decided to make an argument out of something so indefensible. Bizarre.
No, what's bizarre here is your failure to understand simple context. That's on you. Seems to indicate to me a lack of experience. Also, spell-check is your friend..
 
Last edited:
Don't be pedantic. We're not talking about standard computer hardware. We're talking about mostly specialized water based fluid circulation hardware. They are in very difference classes and types of equipment. And yes, the competency of the installer is relevant in all cases.

AIO's are greatly easier to install and require MUCH less experience, thus much lower chance of screwing it up.


No, what's bizarre here is your failure to understand simple context. That's on you. Seems to indicate to me a lack of experience. Also, spell-check is your friend..
Wow I think I'm going to keep this series of posts as a screenshot and share it. I have to ask - are you being intentionally obtuse?

So you are now doubling down (or is it tripling down now?) on your claim that to compare measurable cooling performance levels of a specific AiO vs a specific custom loop, the competency of the person doing the installing is also part of the physical testing (despite being unrelated to the cooling potential of the hardware itself), as opposed to just ensuring both are installed correctly before doing comparative testing, and multiple tests are run to eliminate outliers such as poor tim application or faulty installation.

Please point me to ANY single instance where a cooling hardware performance comparison has included measuring the competency of the installer. Please, just one will do, from any legit source. You can go back the last 15-20 years if you want, which is how long I've been building rigs. I really want to see this, since you're so adamant about it.

Heh.. I can't believe you're actually rolling with this. It's hilarious!

Sad attempt to deflect by criticizing a simple spelling error btw. Also please learn the difference between subjectivity and objectivity while you're at it, and maybe a little about scientific method.
 
Wow I think I'm going to keep this series of posts as a screenshot and share it.
Feel free drama queen.
I have to ask - are you being intentionally obtuse?
Crack a dictionary.
Please point me to ANY single instance where a cooling hardware performance comparison has included measuring the competency of the installer. Please, just one will do, from any legit source. You can go back the last 15-20 years if you want, which is how long I've been building rigs. I really want to see this, since you're so adamant about it.
First point, the onus is not on me to cite COMMON knowledge. You don't give a monkey tools to work on a rocket and expect reasonable results. Second point, I'm not your lackey. Go find your own info.
Heh.. I can't believe you're actually rolling with this. It's hilarious!
That says much more about you than it does me.

And what it says is that you can't debate merit anymore so you're reacting with your pride & ego and making it personal. What a shocker..
 
Feel free drama queen.

Crack a dictionary.

First point, the onus is not on me to cite COMMON knowledge. You don't give a monkey tools to work on a rocket and expect reasonable results. Second point, I'm not your lackey. Go find your own info.

That says much more about you than it does me.

And what it says is that you can't debate merit anymore so you're reacting with your pride & ego and making it personal. What a shocker..
Wow.. I really struck a nerve with you.

Intentionally obtuse. Crack a dictionary indeed:
Obtuse:
: lacking sharpness or quickness of sensibility or intellect : insensitive, stupid
: difficult to comprehend : not clear or precise in thought or expression

Please feel free to Google AiO vs custom loops - there are literally hundreds if not thousands of articles and testing on this very topic, numerous carried out by professionals. You can contact them and tell them they are all wrong if that makes you feel better.

You come across as extremely arrogant, or perhaps you have mental health issues. I'm not sure which it is, but if it's the latter, I apologize for my lack of understanding of your condition. If it's the former, well, you're just being you I guess. A brief look at your history shows this kind of interaction is a common occurrence for you. Maybe that should tell you something. Either way, I am removing myself from this pointless back and forth because it's just silly and a waste of time. Have a nice day.
 
Low quality post by lexluthermiester
@Nereus @lexluthermiester

There are valid points to both arguments. AiO's have made access to water cooling easier. Better? Only if the right components are chosen. Even for an experienced hobbyist, choosing the right pump, the correct connections, the right rads, the correct fans for optimum airflow/noise, etc can make a huge difference. And there's no rulebook, as such. People come to forums for answers. Which I guess is what this discussion is about. But try this - be civil - both of you, please.

I'll not ask again. Make your points with logic and don't deflect.

Thanks.
 
You have not tested every AIO on the market. I do not even see the Eisbaer in your testing and that AIO uses nothing but actual Water cooling parts. We are talking about the best AIOs vs a loop. Modern CPUs are very dense and wattage is not the only contributor to heat. If you mean the Asetek based coolers I agree. Even Thermalright have better and cheaper coolers than any of those units.
The only one close is Swiftech Apogee which is EOL years ago, however that one had the obvious design flaw of putting a DDC pump coupled directly to the cpu block (lol). That one was a real block, albeit they had to compromise with flow performance due to limitations with an actual DDC mounted on top. The eisbaer uses a weak DC-LT pump, low speed and low head pressure, granted its a lot better than the junk in those Asetek based AIO, but it's not a D5 and no where near a DDC.

Please point me to ANY single instance where a cooling hardware performance comparison has included measuring the competency of the installer. Please, just one will do, from any legit source. You can go back the last 15-20 years if you want, which is how long I've been building rigs. I really want to see this, since you're so adamant about it.
I remember back in the day testers used to do 5-10 mounts, compare the quality of each mounting, then average the results to remove the variability of the user from the equations. I wouldn't bother debating, you're just wasting your keystrokes.
 
What is a Custom loop anyway?
 
What is a Custom loop anyway?

It's a highly cost-effective efficient cooling device. A long tube. One end goes in your mouth and the other end over the CPU... the harder you blow the better the CPU temps hehe.
 
Just throwing this out there.. I am using a Thermalright Aqua Elite 360 V3 with different fans.. this thing is really good. I gave AIOs a hard time in the past, with good reason.. but this is a good unit. And it is cheap. Not sure who makes it but they did a good job. It had only been a few months though, time will tell.. but I can say this unit is hands down better than my old H100 with 120x38s. I am running push pull right now, mainly just to get some extra air across my ram and nvme.
 
The problem is, watercooling components are really freaking expensive
IMO the price difference tells you everything you need to know. Neither has killed the other, they're competing in two very distinct markets.

Custom loops are a luxury product. They're prohibitively expensive, but (when done properly) they offer the best performance of any cooling method short of the mad-scientist stuff (chillers and whatnot). They're also a statement piece - having a custom loop, especially with hard tubes, is a flex. Nerds like us know the high cost and effort to make a custom loop, and "normal" people see pretty LEDs and water inside a computer.

On the other hand, AIOs are a mass-market "average joe" product. They prioritize ease of use, and they're made to an accessible price point. That doesn't mean there's anything wrong with AIOs, they just cater to a different set of needs. They can cool much more effectively than a plain ol' air cooler, but without breaking the bank.

Just popped in my head. I would not be against testing it though.

Please, PLEASE do this, and post pictures.
 
I remember back in the day testers used to do 5-10 mounts, compare the quality of each mounting, then average the results to remove the variability of the user from the equations. I wouldn't bother debating, you're just wasting your keystrokes.
Yeah that's true, I mentioned removing outliers, and that's part of the method. :) ..then they compare runs (3DMark etc) for cooling effectiveness. The more runs, the better.
I wouldn't bother debating, you're just wasting your keystrokes.
lol.. I believe that went straight over someones head.. :laugh: ..or maybe my own. :(
 
AIO is nice and easy for me. I really do not want spend time setting up a custom loop for minimal decrease in temps. AIO does the job just fine. The less I spend building my pc the better. I don't want to do the work and mess for that. I don't even like the way it looks either. AIO is fine for me.
 
AIOs are nice, until they fail. I had to replace my TRs AIO with air cooling. God so much better. Plus if an air cooler fails, it's like a fan at most. If it's an AIO, you throw it away.
What a bloody waste.
 
Wow I think I'm going to keep this series of posts as a screenshot and share it. I have to ask - are you being intentionally obtuse?

So you are now doubling down (or is it tripling down now?) on your claim that to compare measurable cooling performance levels of a specific AiO vs a specific custom loop, the competency of the person doing the installing is also part of the physical testing (despite being unrelated to the cooling potential of the hardware itself), as opposed to just ensuring both are installed correctly before doing comparative testing, and multiple tests are run to eliminate outliers such as poor tim application or faulty installation.

Please point me to ANY single instance where a cooling hardware performance comparison has included measuring the competency of the installer. Please, just one will do, from any legit source. You can go back the last 15-20 years if you want, which is how long I've been building rigs. I really want to see this, since you're so adamant about it.

Heh.. I can't believe you're actually rolling with this. It's hilarious!

Sad attempt to deflect by criticizing a simple spelling error btw. Also please learn the difference between subjectivity and objectivity while you're at it, and maybe a little about scientific method.

Before you say something so indefensible yourself...maybe you should think about what you are saying. In the last 3 years:
1) Gamers Nexus has started including pressure maps of the CPU-Cooler interface to demonstrate theoretical performance can be influenced by bad installs.
2) Likewise, their ratings at GN are done before and after they take the computer apart...and several builds have performed different based on who did the cooler install.
3) If you don't want to believe GN...for some reason...then the purchasable brackets for recent Intel CPUs a improve thermal performance by relocating forces on the CPU interface is proof that professional overclockers acknowledge that installation is a consideration.

Now...all of the above hurts me to say. Theory is great...but it's not the full picture. You're welcome to argue...but it is a fact that how you install a CPU cooler can have an immense impact on its performance, and that "using the instructions" is not a viable answer to this performance gap. I...cannot believe that the gauntlet being thrown down literally would take 20 seconds on youtube to verify as a challenge showing your own ignorance...


What you don't see is groups like LTT and some others calling this out... I also know that they screw up results through basic analysis...so maybe one stupidity leads to another? I'm good with all of this if you'd be able to mathematically prove your point, but news flash is that you cannot. The math side says that surface area of contact and thermal properties influence heat transfer rates...which is directly linked to cooling performance. Your assumption on the mathematics is a perfect situation...so your own garbage assumption (left unstated) produces garbage results. GIGO is a thing...and if you don't understand those assumptions it's easy for someone to question them and you to not be able to comprehend why they are so...fundamentally required to be consistent only if you are working in theory land. In practical land, where we live, a theory is only as good as its practical use...and the news flash here is that installation is 100% directly linked to performance.
 
Before you say something so indefensible yourself...maybe you should think about what you are saying. In the last 3 years:
1) Gamers Nexus has started including pressure maps of the CPU-Cooler interface to demonstrate theoretical performance can be influenced by bad installs.
2) Likewise, their ratings at GN are done before and after they take the computer apart...and several builds have performed different based on who did the cooler install.
3) If you don't want to believe GN...for some reason...then the purchasable brackets for recent Intel CPUs a improve thermal performance by relocating forces on the CPU interface is proof that professional overclockers acknowledge that installation is a consideration.

Now...all of the above hurts me to say. Theory is great...but it's not the full picture. You're welcome to argue...but it is a fact that how you install a CPU cooler can have an immense impact on its performance, and that "using the instructions" is not a viable answer to this performance gap. I...cannot believe that the gauntlet being thrown down literally would take 20 seconds on youtube to verify as a challenge showing your own ignorance...


What you don't see is groups like LTT and some others calling this out... I also know that they screw up results through basic analysis...so maybe one stupidity leads to another? I'm good with all of this if you'd be able to mathematically prove your point, but news flash is that you cannot. The math side says that surface area of contact and thermal properties influence heat transfer rates...which is directly linked to cooling performance. Your assumption on the mathematics is a perfect situation...so your own garbage assumption (left unstated) produces garbage results. GIGO is a thing...and if you don't understand those assumptions it's easy for someone to question them and you to not be able to comprehend why they are so...fundamentally required to be consistent only if you are working in theory land. In practical land, where we live, a theory is only as good as its practical use...and the news flash here is that installation is 100% directly linked to performance.
Maybe you misunderstood me, or didn't read all the posts? For sure installation can be an issue - tim not spreading evenly, block not screwed down properly, among other issues. Looking at the tim after testing, using thermal cameras etc has been going on a LOT longer than just 3 years - I am well aware of this. My original point was that - once all these variables have been eliminated (as much as reasonably possible, which is why multiple tests and multiple installations are run, not just one) - custom loops generally perform better than AiO. That was my only point, which is my opinion from years of personal experience and years of watching relative professionals carry out testing. AiO's are getting better for sure, and are way more convenient, and less likely to be screwed up with installation - I have not been arguing against any of that - but once they are reasonably assured to be good installs, loops generally perform better at cooling than AiO (more definitively under high load testing) - assuming both are using high quality components. Seems to be a bunch of butthurt AiO users here who don't like to hear that and have repeatedly taken what I've said either out of context or just twisted it intentionally. Get over it already. If you think AiO's cool better than custom loops then go ahead and make your argument rather than attacking me! ffs.... I never realized this place was so toxic.

Fwiw, I have my main rig on full loop, and a smaller mini-ITX that uses AiO cooling, so I am not anti-AiO or anything like that. Just stating my opinion. Sad that it seems to be met with such anger and disdain. My mistake was reacting to those individual(s).

AIOs are nice, until they fail. I had to replace my TRs AIO with air cooling. God so much better. Plus if an air cooler fails, it's like a fan at most. If it's an AIO, you throw it away.
What a bloody waste.
True that. Although I recall EK(?) and EVGA both experimented years ago with AiOs that had quick disconnect fittings. Didn't seem to take off though.
 
Last edited:
I never realized this place was so toxic.

It's not toxic. People are. And we do a lot to keep it civil, but when you use terms like 'butthurt' you're just asking for kickback. Similarly, others will argue their point with no intention of trying to see the shades of grey. Usually, what you have is one person on one side with a strong opinion, and another on the opposite side. In those cases, in any forum, you're not going to get an easy ride.

On the flip side, the arguments I'm seeing being discussed here are just ridiculous. It's a simple question to ask - have AiO's killed custom loops? I think there's a massive difference in audience. I used to use custom loops but my preference was for quiet. AiO's weren't so great back then. But now? If I wanted to go water, I'd probably use an AiO because I can't be arsed with the complexity of a custom loop. Tie that with the perceived costs of a good custom set-up, and it's easier to see the hit AiO's have had on custom loops. That's my view, anyway.

I'd not attack either side on either preference, but in my mind, for me, it's absolutely not worth the hassle. If I had more spare time, and PC builds were a hobby, sure, I'd probably go custom. For one, you can make some beautiful looking set ups that AiO's (subjectively speaking) can't match. As far as cooling properties go, a custom loop (installed correctly) with the same water volume and cooling capacity offers nothing over an AiO. But then again, I can always buy a 1080mm custom radiator, 9 fans, and have a marginal benefit over an AiO. I can also (as has been done by some nutcases) place custom loop external to the PC. Even outside. That, I suppose is a major advantage of a custom loop - the expandibility of it. Until of course, you buy a hybrid AiO with disconnects... (but that's really a custom set up).
 
It's not toxic. People are.
Heh ok, true enough. :)
..but when you use terms like 'butthurt' you're just asking for kickback.
True again. As I also stated, "My mistake was reacting to those individual(s)". My bad. I accept that.

Much of the rest of your points are valid, but again, I've already agreed about convenience, expense, etc... All I intended to say from the start was regarding thermal cooling performance and nothing else. Even you sort of deflecting into other aspects here which I, again, don't disagree with and never did. ¯\_(ツ)_/¯

I'll concede one point though; if you used the same block on both systems, and the same radiator, same fans, and eliminated all other variables as much as possible (installation errors, tim application etc.), they are going to be so close as to be negligible. You're still going to have a very small advantage with a custom loop because of the pumps that are used in most AiO compared to a D5 which is pretty much standard in a loop. Logically, if you had the same pump, then they'd perform the same (or close enough to be negligible). To reiterate, this is not talking about performance for day to day usage, this is stress testing on high load for extended periods, which is how these things are usually compared.

FWIW, after EVGA basically ended itself, I've lost much of my enthusiasm for what has been a personal hobby for many years. Quite possibly I will go AiO only in the future, because building and maintaining a custom loop is a bit of a pita once you've lost that passion for it. I just don't have the same desire to run 3DMark runs for hours on end just trying to tweak a few more points out of my system. Life goes on.

I certainly regret posting in this thread though, lol. I'll be careful about what I post in future. Thanks for not going all DEFCON1 and just deleting everything and swinging the ban hammer, but instead talking rationally, that's appreciated - sign of a good mod. Cheers.
 
AIOs are nice, until they fail. I had to replace my TRs AIO with air cooling. God so much better. Plus if an air cooler fails, it's like a fan at most. If it's an AIO, you throw it away.
What a bloody waste.
Just because you had failure doesn't mean someone else will.

i've had my for 5 years now and still going strong.

It's not toxic. People are. And we do a lot to keep it civil, but when you use terms like 'butthurt' you're just asking for kickback. Similarly, others will argue their point with no intention of trying to see the shades of grey. Usually, what you have is one person on one side with a strong opinion, and another on the opposite side. In those cases, in any forum, you're not going to get an easy ride.

On the flip side, the arguments I'm seeing being discussed here are just ridiculous. It's a simple question to ask - have AiO's killed custom loops? I think there's a massive difference in audience. I used to use custom loops but my preference was for quiet. AiO's weren't so great back then. But now? If I wanted to go water, I'd probably use an AiO because I can't be arsed with the complexity of a custom loop. Tie that with the perceived costs of a good custom set-up, and it's easier to see the hit AiO's have had on custom loops. That's my view, anyway.

I'd not attack either side on either preference, but in my mind, for me, it's absolutely not worth the hassle. If I had more spare time, and PC builds were a hobby, sure, I'd probably go custom. For one, you can make some beautiful looking set ups that AiO's (subjectively speaking) can't match. As far as cooling properties go, a custom loop (installed correctly) with the same water volume and cooling capacity offers nothing over an AiO. But then again, I can always buy a 1080mm custom radiator, 9 fans, and have a marginal benefit over an AiO. I can also (as has been done by some nutcases) place custom loop external to the PC. Even outside. That, I suppose is a major advantage of a custom loop - the expandibility of it. Until of course, you buy a hybrid AiO with disconnects... (but that's really a custom set up).
1000%

Some people don't notice if you post aggressively you will get an aggressive response. And a lot of people seem to forget other people are allowed to have a different opinions. There are certain ethics that have to be learned when on a forum on how to communicate. Your points on AIO vs custom also spot on.
 
Last edited:
It really depends on how you use an AIO, if you give it an easy life with a low tdp it should last awhile. If you are using a clocked up fire breathing dragon running F@H or something all the time, you should be able to get at least a year out of it
 
It really depends on how you use an AIO, if you give it an easy life with a low tdp it should last awhile. If you are using a clocked up fire breathing dragon running F@H or something all the time, you should be able to get at least a year out of it
Agreed if you are rocking a 12900K,13900K,14900K I would go custom loop for that over an AIO.

Any cpu's that are sub 150 Watts I think AIO is a better option.
 
I have had multiple AIOs. Not a problem with one of them. I game normally. I have a 7900X now with an AIO. 0 problems.
 
I have had multiple AIOs. Not a problem with one of them. I game normally. I have a 7900X now with an AIO. 0 problems.
I run my 5900X @ 260w and it’s fine with air and AIO, though a little better with AIO in my current config.. right now I am running to see how it fares the test of time.. though the majority of the time my 5900X is being run nicely, but quickly..
 
If you are using a clocked up fire breathing dragon running F@H or something all the time, you should be able to get at least a year out of it

Wow, I'm starting to get close to 20 years out of one of my D5 pumps, it was called a Swifttech MCP-655 back then. Was purchased around 2007 and still going strong.

1707330738221.png
 
Back
Top