Question Will NZXT Kraken Elite 360 adequately cool a TR5/SP6 socket CPU?

Mar 18, 2024
28
8
35
I am considering switching from air-cooling to AIO for my AMD 7965WX. But I am wondering if it will fit properly and adequately cool a CPU for the TR5/SP6 socket.

I have been looking at the NZXT Kraken Elite 360mm and am wondering if it will work with my CPU. The product description on their site says:
Compatible Socket & CPU: AMD Socket AM5, AM4, sTRX4*, TR4* (*Threadripper bracket not included) AMD Ryzen 9 / Ryzen 7 / Ryzen 5 / Ryzen 3 / Threadripper"
... but I can't find the "Threadripper bracket" listed anywhere on their site, and it only says TR4, not TR5.

Has anyone put these on a TR5 and have experience with effectiveness? Also, which mounting hardware do I need to get? (I'm using WRX90E-SAGE motherboard if that matters)

I am using it in a particularly high-heat environment with 2 RTX 4090s jammed in the case along with the CPU, so I really need to make sure it's providing max thermal performance.
 
Mar 18, 2024
28
8
35
I wouldn't go for an AIO since you the CPU's cooling block won't cover the entire IHS of a ThreadRipper.
The problem that I'm running into is that the only air coolers I've been able to find that actually cover the whole TR5 heat spreader are the Noctua U14S TR5-SP6 and the Noctua NH-D9 TR5-SP6 4U.

The U14S is that it is in the wrong orientation for desktop cases (perpindicular to long axis of the CPU heat spreader) so that it pulls air from bottom to top, using hot air from the two 450W GPUs below as intake for the CPU.

Meanwhile, the NH-D9 has the correct orientation (front --> back), but it has tiny 92mm fans and a MUCH smaller cooling block, and from the thermal benchmarks I've seen, it runs significantly hotter than the U14S.

Are there really no good quality AIO systems that will cover the entire IHS of a TR5 CPU?
 
Last edited:
Mar 18, 2024
28
8
35
What about Ice Giant's TR cooler?
Yeah, I saw this one, but it also only lists TR4 (not TR5/SP6) so I am assuming it won't cover the entire IHS.

Contact them and ask how to get it if you buy the cooler.
Yeah, I had sent their support an email through their website, and got a response back that says that I won't hear back until Tuesday or Wednesday at earliest. I doubt it will cover the entire IHS though, because they only list TR4 socket on list of compatible CPUs.
 
Last edited:
  • Like
Reactions: ReveurGAM

ReveurGAM

Commendable
Sep 28, 2022
396
21
1,695
Yeah, I saw this one, but it also only lists TR4 (not TR5/SP6) so I am assuming it won't cover the entire IHS.


Yeah, I had sent their support an email through their website, and got a response back that says that I won't hear back until Tuesday or Wednesday at earliest. I doubt it will cover the entire IHS though, because they only list TR4 socket on list of compatible CPUs.
Well, who knows? Maybe it does. Maybe the Ice Giant does. I wish I had more suggestions for you.
 
Mar 18, 2024
28
8
35
Well, who knows? Maybe it does. Maybe the Ice Giant does. I wish I had more suggestions for you.
NZXT got back to me and says they don't sell mounting brackets, and that I'd need to talk to AMD.

I already made the mistake of buying one TR4 cooler (Arctic Freezer 4U-M), because it said it was "compatible", and it was possible to attach, but it didn't cover the whole IHS. I ended up returning it and going with a Noctua TR5-SP6 cooler because it's the only one that actually covers the CPU.

So many of these cooler manufacturers using dishonest marketing to say their inadequate TR4 cooling solutions are "compatible" with TR5/SP6 socket just because it's possible to screw them on. These processors are HOT (especially 24 core and higher), and really need something that fully covers the CPU heat spreader to maximize thermal transfer, and these TR4 coolers don't cut it ....

Apparently Silverstone will be releasing an AIO system for TR5/SP6 this June.

Until then, it seems like air cooled (Noctua) is the only option if you want to cover the entire IHS.
 
Last edited:
  • Like
Reactions: ReveurGAM
You don't actually need to cover the entire IHS on these CPUs just where the cores and IO die are. If you look at what cooling is used in most of the press reviews they're standard asetek AIOs. The Arctic cooler you returned is pretty much the best air cooler option right now as the Noctua perform no better and cost significantly more.

FWIW the Silverstone AIO will cost a lot when it comes out as it'll be the same design as their LGA 4677 one which is >$300.
 
Mar 18, 2024
28
8
35
You don't actually need to cover the entire IHS on these CPUs just where the cores and IO die are. If you look at what cooling is used in most of the press reviews they're standard asetek AIOs. The Arctic cooler you returned is pretty much the best air cooler option right now as the Noctua perform no better and cost significantly more.

FWIW the Silverstone AIO will cost a lot when it comes out as it'll be the same design as their LGA 4677 one which is >$300.
Yes, it is generally my understanding that you don't *need* to cover the entire IHS in many cases. But all other things being equal (size of cooling tower, speed of fans, etc) covering the entire IHS will conduct more heat, right?

I understand that under "normal" loads, this is a non-issue and that you just need to have the cores/die covered. But I am running two air-cooled RTX 4090s in the same case (as well as 8 sticks of DDR5, 3 SSDs, and 2 HDDs), so there is a lot more heat to deal with than normal so even shaving off just a few degrees C would be helpful.

Also, on a more subjective note, the construction of the Noctua coolers feels much more solid/sturdy. The Arctic cooler felt cheaper, and like it was manufactured with looser tolerances.

But I am willing to accept that maybe I am wrong about the Arctic cooler (although at this point, it's too late to change) ... I will likely consider AIO when it comes out though. Right now, while the spring weather is cooler here, my thermals seem fine with the Noctua, but I'm worried that as the summer picks up and indoor temps climb that thermals might be more of an issue. So the june release of the Silverstone AIO would be just in time :)

Is it not possible to rotate the cooler? The tr3-sp4 could be mounted front-back.
No, it is not possible to rotate the Noctua U14S to be front-back. Only the NH-D9 mounts this way.
 
  • Like
Reactions: 35below0

35below0

Respectable
Jan 3, 2024
1,727
744
2,090
Also, on a more subjective note, the construction of the Noctua coolers feels much more solid/sturdy. The Arctic cooler felt cheaper, and like it was manufactured with looser tolerances.
Funny you should say this. I was installing Noctuas and Arctics just last week. I felt the same way. Arctic came in a cheaper box too.

But the performance? It's there. Arctic just doesn't feel premium and maybe isn't as fine and quiet as Noctua but it's good.

Btw, the Noctua redux line also feels cheaper. Well, it IS cheaper so...

What case are you using? If you don't mind the price, the Fractal Torrent is hard to beat. Most of it's air hits the GPU(s). It might get loud if under heavy load.
The XL version steps up to 180mm front intake fans...
 
  • Like
Reactions: jferments
Mar 18, 2024
28
8
35
What case are you using? If you don't mind the price, the Fractal Torrent is hard to beat. Most of it's air hits the GPU(s). It might get loud if under heavy load.
The XL version steps up to 180mm front intake fans...
I am using a Fractal Meshify 2XL. I am very satisfied with it. My only complaint is that while it technically does work with SSI-EEB motherboards (I'm using WRX90E-SAGE), the cable routing ports on the right-hand side back panel are covered up by the motherboard, so my cable management was a bit messier.

I've currently got 5 x 140mm 1000RPM chassis fans (3 front intake and 1 rear + 1top exhaust). I've got two 140mm 3000rpm Noctua fans on the way though that I'm going to use for front intake (to blow over the CPU and top of the GPUs), moving two of the 1000RPM fans to the top for exhaust.
 
  • Like
Reactions: 35below0
Mar 18, 2024
28
8
35
Think twice about this. Having a top fan too close to the intake fans will rob the case of some of the air. It will get sucked up and out before it's had a chance to cool anything.

Just a thought.
Thanks for the feedback. I was thinking that with the top fans being 1000rpm and the two upper front fans being 3000rpm that this would be less of an issue ... but yeah, maybe I'll just do two in the top back and leave the top slot that is closest to the front open 🤔

I'm still within return period for the 1000rpm fans, so I could send one back and buy myself a lunch
 
Yes, it is generally my understanding that you don't *need* to cover the entire IHS in many cases. But all other things being equal (size of cooling tower, speed of fans, etc) covering the entire IHS will conduct more heat, right?
Doesn't even remotely matter so long as there's enough surface to cover the important parts. If you were running a 64 core+ part perhaps it'd make a difference, but even then it doesn't really seem to. Now I wouldn't try to use a regular desktop air cooler on one and with AIOs it would depend on which cold plate
Also, on a more subjective note, the construction of the Noctua coolers feels much more solid/sturdy. The Arctic cooler felt cheaper, and like it was manufactured with looser tolerances.
Oh I don't think that's subjective at all. Arctic's coolers are designed to be mass produced for datacenter/workstation OEM type use. That's why even though the Arctic has more mass it's the same. performance rather than being better than Noctua's.
But I am willing to accept that maybe I am wrong about the Arctic cooler (although at this point, it's too late to change) ... I will likely consider AIO when it comes out though. Right now, while the spring weather is cooler here, my thermals seem fine with the Noctua, but I'm worried that as the summer picks up and indoor temps climb that thermals might be more of an issue. So the june release of the Silverstone AIO would be just in time :)
Keep in mind getting heat out of your case is the biggest part when there are hot ambient temps. I do think the AIO would do a better job as you'd then have the CPU heat mostly at the top and would have more open airflow to exhaust GPU heat. That's why even though I cringed at the thought of spending so much on a non custom loop cooler Silverstone's AIO was on my list when I was contemplating a SPR workstation build.
 
  • Like
Reactions: jferments

ReveurGAM

Commendable
Sep 28, 2022
396
21
1,695
Thanks for the feedback. I was thinking that with the top fans being 1000rpm and the two upper front fans being 3000rpm that this would be less of an issue ... but yeah, maybe I'll just do two in the top back and leave the top slot that is closest to the front open 🤔
There is a way to accomplish this. Simply put a barrier between the top- front and front- top fans
 
  • Like
Reactions: 35below0

ReveurGAM

Commendable
Sep 28, 2022
396
21
1,695
Doesn't even remotely matter so long as there's enough surface to cover the important parts. If you were running a 64 core+ part perhaps it'd make a difference, but even then it doesn't really seem to. Now I wouldn't try to use a regular desktop air cooler on one and with AIOs it would depend on which cold plate

Oh I don't think that's subjective at all. Arctic's coolers are designed to be mass produced for datacenter/workstation OEM type use. That's why even though the Arctic has more mass it's the same. performance rather than being better than Noctua's.

Keep in mind getting heat out of your case is the biggest part when there are hot ambient temps. I do think the AIO would do a better job as you'd then have the CPU heat mostly at the top and would have more open airflow to exhaust GPU heat. That's why even though I cringed at the thought of spending so much on a non custom loop cooler Silverstone's AIO was on my list when I was contemplating a SPR workstation build.
Top mount is the least valuable position for an AIO. Testing that I and a peer have done demonstrate that if the front is somewhat blocked (e.g. glass plate on the 5000X) and you can side mount, it will be the best option. If there is no significant front barrier, it will be better than side mount. Top mount is almost always the worst of the 3, although there will always be exceptions. This can be somewhat mitigated through clever fan placement.

As for push vs pull, tubes up vs down, GC vertical vs mobo mount, and intake vs exhaust, they are not individually that important but, combined, can nerf certain options.

The only exception I can think of is if the rad is outside the case.

Also, make sure to be careful not to have the pump at the top of your loop... some rads come with the pump in they them or nearby on the tubes, so that makes it impossible to safely use them in certain positions.
 
Top mount is the least valuable position for an AIO. Testing that I and a peer have done demonstrate that if the front is somewhat blocked (e.g. glass plate on the 5000X) and you can side mount, it will be the best option. If there is no significant front barrier, it will be better than side mount. Top mount is almost always the worst of the 3, although there will always be exceptions. This can be somewhat mitigated through clever fan placement.
No you're ignoring the system this person is building and are absolutely wrong. They have 2x RTX 4090s in a case with only front intake. It's much smarter to feed those clean air than it is to get optimal CPU cooling.
 

ReveurGAM

Commendable
Sep 28, 2022
396
21
1,695
No you're ignoring the system this person is building and are absolutely wrong. They have 2x RTX 4090s in a case with only front intake. It's much smarter to feed those clean air than it is to get optimal CPU cooling.
Any time I see someone make a blanket statement, it's 98% certain that they are either partially or fully wrong, or they are lying. Your statement is probably wrong, so I'll just assume you made an error.
  1. I stated that based on testing conducted by myself and another person, the top is the worst.
  2. I stated that there exceptions to these results based on various factors. This statement is eminently applicable and is the main thing that demonstrates the fallacy of your spurious claim.
  3. Two mounted GCs would be best served by fans on the front because they will feed air towards the GCs, especially if the OP users ducting to direct air to them.
  4. Additional fans that are strategically placed will provide extra support to the GCs.
  5. The CPU roasts far more readily than GCs, and especially because of GCs. The CPU is the first priority.
  6. Finally, according to you: "in a case with only front intake. " that statement by itself demonstrated that you are wrong because, if there is ONLY front intake, there is no way to top mount!
The OP hasn't provided their system specs, so I can't and didn't give more specific recommendations. If you have facts to share, then please share them instead of making half- baked claims.
 
Mar 18, 2024
28
8
35
I know very little about AIO systems, so I am watching this conversation mostly from a place of trying to learn.

But as far as my current system, I want to confirm that the GPUs are struggling way less to cool themselves than the CPU even though they are sandwiched right up next to each other. Even when both 4090s have their VRAM fully loaded and doing heavy workloads (large LLM models, stable diffusion, etc) they don't even get close to max temps. Even the upper card rarely breaks 75C (it generally runs about 5-10C hotter than the one below). But the CPU (which is above them and taking in their hot air as intake) definitely gets hotter and closer to thermal limits (only when CPU and GPUs are all running at full load). This is why I was going to position the two 3000 RPM fans to blow over the CPU and tops of the GPUs.

But I am very curious to understand why a top mounted AIO wouldn't be best. This seems very counterintuitive to me, because I would think that venting heat from CPU cooler out the top would be most efficient, since heat rises. Can you explain why you think your testing showed this to be the WORST case?
 
Any time I see someone make a blanket statement, it's 98% certain that they are either partially or fully wrong, or they are lying. Your statement is probably wrong, so I'll just assume you made an error.
  1. I stated that based on testing conducted by myself and another person, the top is the worst.
  2. I stated that there exceptions to these results based on various factors. This statement is eminently applicable and is the main thing that demonstrates the fallacy of your spurious claim.
  3. Two mounted GCs would be best served by fans on the front because they will feed air towards the GCs, especially if the OP users ducting to direct air to them.
  4. Additional fans that are strategically placed will provide extra support to the GCs.
  5. The CPU roasts far more readily than GCs, and especially because of GCs. The CPU is the first priority.
  6. Finally, according to you: "in a case with only front intake. " that statement by itself demonstrated that you are wrong because, if there is ONLY front intake, there is no way to top mount!
The OP hasn't provided their system specs, so I can't and didn't give more specific recommendations. If you have facts to share, then please share them instead of making half- baked claims.
You obviously didn't read this thread where they said everything about their system. Perhaps before typing walls of text you should try reading first.
 

ReveurGAM

Commendable
Sep 28, 2022
396
21
1,695
I know very little about AIO systems, so I am watching this conversation mostly from a place of trying to learn.

But as far as my current system, I want to confirm that the GPUs are struggling way less to cool themselves than the CPU even though they are sandwiched right up next to each other. Even when both 4090s have their VRAM fully loaded and doing heavy workloads (large LLM models, stable diffusion, etc) they don't even get close to max temps. Even the upper card rarely breaks 75C (it generally runs about 5-10C hotter than the one below). But the CPU (which is above them and taking in their hot air as intake) definitely gets hotter and closer to thermal limits (only when CPU and GPUs are all running at full load). This is why I was going to position the two 3000 RPM fans to blow over the CPU and tops of the GPUs.

But I am very curious to understand why a top mounted AIO wouldn't be best. This seems very counterintuitive to me, because I would think that venting heat from CPU cooler out the top would be most efficient, since heat rises. Can you explain why you think your testing showed this to be the WORST case?
I'm still analyzing the data, and i haven't yet seen anything that helps me to understand why. these results have already been achieved by testers before our separate tests, and MSI points to front mount.

If you want to see more, you'll need to PM me because I'm not allowed to link to it on penalty of being banned.

You obviously didn't read this thread where they said everything about their system. Perhaps before typing walls of text you should try reading first.
Perhaps before making a useless response without actually proving yourself correct, you should read first. 😆 🤣 😂 😹 😆 🤣 😂 😹

FWIW, I probably did read it but forgot. That doesn't even matter, though. The OP already verified my point.