NH-U14S-TR4 running hot on stock 1920X

[HTC]

The motherboard is the Asus ROG Strix X399-E Gaming. It's listed as fully compatible (as is every other TR4 motherboard except for the Zenith Extreme, which has a PCIe clearance issue).

Suppose I might get a second fan somewhere down the line then. Does it really make that much of a difference on a mid-sized heatsink like this? Outside of once sticking two fans on a Hyper 212 Evo (where it didn't even shift temps 1 degree), my only real experience with push-pull setups is for water cooling, where it makes no difference at all unless you have a very thick radiator and really need the static pressure. Of course, a heatsink like this is far thicker than most radiators, so I guess that kind of explains why it might work better?

I'll look into getting a pic of the cooler in the case the next time I bother to pull it out from under the desk

I suspected as much about the compatibility but, because i knew about that Zenith board, i'd thought i'd ask: i'm not familiar with all TR4 boards, after all.

As evidenced here, it does make quite a difference: nearly 7 degrees with a 2990WX with PB2 enabled.

 

[Valantar]

I suspected as much about the compatibility but, because i knew about that Zenith board, i'd thought i'd ask: i'm not familiar with all TR4 boards, after all.

As evidenced here, it does make quite a difference: nearly 7 degrees with a 2990WX with PB2 enabled.

I kind of wish they'd run a 1-vs-2-fan comparison on the 2950X. Even if the 1920X lacks PB2, and thus sticks to base clocks when loads exceed ... 2 cores, I think?, that shouldn't matter much given that I've seen 182W power draws at 3.5GHz on this unit, at least as reported by HWMonitor. (Btw, it's interesting to see just how much power Infinity Fabric, memory controllers and the like contributes - running OCCT, the cores never exceeded 116W, with the remaining ~70W thus being drawn by these other parts. From what I've read (I think AnandTech did a piece on this once), Intel's mesh interconnect is very much in the same league of power draw.) Also, it seems like the better distributed heat load (4 dice vs. 2) on the 2990WX actually helps the U14S perform a bit better, given that it was struggling to maintain 180W on the 2950X. I get that a second fan helps, but a 39% increase in cooling potential (180W at 100% fan speed to 250W at 100% fan speed) from that alone seems unrealistic to me. Suppose I should run some before-and-after testing when I get the second fan

 

[RealNeil]

I kind of wish they'd run a 1-vs-2-fan comparison on the 2950X.

I kind of wish you'd run a 1-vs-2-fan comparison on the 2950X and your cooler.

I have a feeling that one of two things would happen.
1. you would be pleasantly surprised at your cooler running CPU.
2. You would discover that your air cooler (no matter the brand, configuration, and quality) is not up to the task for cooling that CPU.

You originally said that the AIO cooler you had on it failed after a short time. You Said: (wouldn't trust a replacement AIO given that the first one failed after less than a year ...)

I imagine that it was doing a proper job cooling the 2950X or you would have mentioned that besides its crappy reliability. I propose that (if the air cooler can't do the deed) you get a better AIO for your system. I'm using them on my i9-7900X, and my two i7-8700K systems. If you get good ones, they just work and they're usually reliable.
The only AIO that ever failed on me was a Silverstone brand 240mm, and now I just don't buy that brand anymore.

I tried a massive Air Cooler on my i9-7900X system first. It was a CoolerMaster Master Air Maker-8. It was completely capable with my X99 i7-6850 system overclocked. But the 10 core i9-7900X was too much for it.

I guess that what I'm saying is not to limit your options just because of an uncharacteristic AIO failure.

Good luck with it.

 

[Deleted member 178884]

I guess that what I'm saying is not to limit your options just because of an uncharacteristic AIO failure.

I agree, people whine about the silent loop 280mm having "noise issues" yet it's held up a 5ghz delidded 7740x and my previous 6600k before x299 was even out - it's been running flawlessly and has lasted years and still runs without fault. Hell even a alphacool eisbaer 360mm I bought runs flawlessly and that was used with oxidation on the copper plate, no issues with that either on my 6600k rig.

 

[HTC]

I kind of wish you'd run a 1-vs-2-fan comparison on the 2950X and your cooler.

I have a feeling that one of two things would happen.
1. you would be pleasantly surprised at your cooler running CPU.
2. You would discover that your air cooler (no matter the brand, configuration, and quality) is not up to the task for cooling that CPU.

You originally said that the AIO cooler you had on it failed after a short time. You Said: (wouldn't trust a replacement AIO given that the first one failed after less than a year ...)

I imagine that it was doing a proper job cooling the 2950X or you would have mentioned that besides its crappy reliability. I propose that (if the air cooler can't do the deed) you get a better AIO for your system. I'm using them on my i9-7900X, and my two i7-8700K systems. If you get good ones, they just work and they're usually reliable.
The only AIO that ever failed on me was a Silverstone brand 240mm, and now I just don't buy that brand anymore.

I tried a massive Air Cooler on my i9-7900X system first. It was a CoolerMaster Master Air Maker-8. It was completely capable with my X99 i7-6850 system overclocked. But the 10 core i9-7900X was too much for it.

I guess that what I'm saying is not to limit your options just because of an uncharacteristic AIO failure.

Good luck with it.

You are confused, dude: OP's girlfriend has an 1920X in her PC. The 2950X is from the review i posted earlier: if the cooler is capable of cooling the 2950X, it should be capable of cooling an 1920X since it has 4 less cores, no?

 

[Valantar]

I kind of wish you'd run a 1-vs-2-fan comparison on the 2950X and your cooler.

Well, there's "only" a 1920X in this PC, but as I said, I suppose I should run some before/after testing when I get around to getting the second fan.

I have a feeling that one of two things would happen.
1. you would be pleasantly surprised at your cooler running CPU.
2. You would discover that your air cooler (no matter the brand, configuration, and quality) is not up to the task for cooling that CPU.

Entirely possible, though that would be grounds for an RMA, and I'd find some other solution.

You originally said that the AIO cooler you had on it failed after a short time. You Said: (wouldn't trust a replacement AIO given that the first one failed after less than a year ...)

I imagine that it was doing a proper job cooling the 2950X or you would have mentioned that besides its crappy reliability. I propose that (if the air cooler can't do the deed) you get a better AIO for your system. I'm using them on my i9-7900X, and my two i7-8700K systems. If you get good ones, they just work and they're usually reliable.
The only AIO that ever failed on me was a Silverstone brand 240mm, and now I just don't buy that brand anymore.

I only did some cursory thermal testing on the AIO when the PC was built (assumed it would be entirely okay given its "500W" rating, and everything worked as it should), but it's been fine* running multi-hour video renders for a while, with no noticeable slowdowns, so I assume temps have been a-ok. The issue is this: there (still!) aren't many AIOs for Threadripper out there (and there sure weren't a year ago), and I'm not going the Asetek-with-a-bracket route, as there's plenty of data showing how that's a very poor solution (insufficient IHS coverage, bad temps, lots of noise. The * next to "fine" above here is due to noise: the fans on the Enermax LiqTech 240 are great in terms of airflow and pressure, but hoooo boy are they noisy. Not what you want running multi-hour renders when the office is also your living room. No amount of fan curve tuning could fix it, sadly. I was considering getting a pair of the new NF-A12x25s for it, but that really shouldn't be necessary for an AIO this expensive.

I guess that what I'm saying is not to limit your options just because of an uncharacteristic AIO failure.

Good luck with it.

The failure might have been a freak occurrence, but I don't trust that given how it failed. If this had been a dead pump, I'd have chalked it down to "oh, we got a bad one" and had it replaced. Instead, it clogged (or, at least, there was no flow) despite the pump still running at full speed (both as reported to the motherboard and audible when connected to a 12V source without the PC running). Which means either the impeller blades all broke (yeah, that doesn't happen), there was growth in the tubes, or it corroded. My money's on the last option, seeing how this is an aluminium radiator with a huge honking copper cold plate next to the CPU. I trusted that Enermax knew more or less what they were doing even with mixing metals, but apparently they didn't.

In other words: the change was prompted by general unhappiness with the noise level of the cooler, plus the fact that the failure most likely is due to an inherent trait of the design. Given that having it replaced took three damn weeks (which really isn't acceptable given the use case for this PC), I wasn't taking the chance on the same thing happening in another year. If this was my personal PC and uptime/stability wasn't crucial, I wouldn't mind tinkering with it when needed, but we can't have this failing willy-nilly. Slightly less performance is far preferable to multi-week outages. Not to mention that the U14S only has one part that can fail: the fan, which is easily replaceable with off-the-shelf parts. And comes with ... what, a 6-year warranty? Even if it struggles a bit to keep up with the heat load, it's a far more suitable solution. I'd likely have gone for that in the first place if it had launched by the time the PC was built.

 

[HTC]

The failure might have been a freak occurrence, but I don't trust that given how it failed. If this had been a dead pump, I'd have chalked it down to "oh, we got a bad one" and had it replaced. Instead, it clogged (or, at least, there was no flow) despite the pump still running at full speed (both as reported to the motherboard and audible when connected to a 12V source without the PC running). Which means either the impeller blades all broke (yeah, that doesn't happen), there was growth in the tubes, or it corroded. My money's on the last option, seeing how this is an aluminium radiator with a huge honking copper cold plate next to the CPU. I trusted that Enermax knew more or less what they were doing even with mixing metals, but apparently they didn't

If it looked like this after the failure, it wasn't a "freak occurrence".

The dude in that post had four of these Enermax coolers in his systems and three of them died within a year.

 

[Valantar]

If it looked like this after the failure, it wasn't a "freak occurrence".

The dude in that post had four of these Enermax coolers in his systems and three of them died within a year.

Yeah, that's pretty much what I was expecting. I didn't take it apart (would likely have made the RMA even more of a hassle), but I wouldn't be surprised at all if it looked like that. Flow was low enough that even with the CPU thermal throttling HARD, the tubes were room temp 10-15cm from the block. In part it was kind of funny - it showed me that if you have a 12c24t CPU, Windows 10 runs surprisingly okay at 550MHz Also, the thermal throttling/protection setup in the Threadripper series seems on point. Never once shut down or crashed, just throttled down to a crawl (all the while the fans were running wild at ~2300rpm). Kind of makes me wonder what would happen if I did nothing more than point a fan at the IHS and turn the PC on. Not willing to risk that though

Still, I wonder what Enermax did wrong here, given that there seem to be plenty of "standard" asetek-based AIOs with copper cold plates and aluminum rads. Different alloys, perhaps? They're not usually nickel-plated, at least. Or maybe it was some anti-corrosion additive they skipped?

 

[HTC]

Still, I wonder what Enermax did wrong here, given that there seem to be plenty of "standard" asetek-based AIOs with copper cold plates and aluminum rads. Different alloys, perhaps? They're not usually nickel-plated, at least. Or maybe it was some anti-corrosion additive they skipped?

Or added that is making the difference.

On another note, and following the suggestion of @RealNeil , if you do end up getting a 2nd fan, please try and test not only with one VS two fans but also one fan VS the other fan. That way, you can confirm / deny an issue with the fan being lower speed then advertised and, if indeed this is the case, you could try and explain the issue to Noctua and send only the original fan to RMA rather than the whole cooler.

 

[Valantar]

Or added that is making the difference.

On another note, and following the suggestion of @RealNeil , if you do end up getting a 2nd fan, please try and test not only with one VS two fans but also one fan VS the other fan. That way, you can confirm / deny an issue with the fan being lower speed then advertised and, if indeed this is the case, you could try and explain the issue to Noctua and send only the original fan to RMA rather than the whole cooler.

That's a good idea, guess I was reading too quickly to pick up that nuance. Will definitely do that one I get a second fan. For now, I have to pester the store to send me a receipt for the swap, and ask them what happened to the price differential between the two coolers (we paid around 50% more for the Enermax than what the Noctua sells for, so they still owe us something there). If I can convince them to send me a second fan I'll be happy to leave them the rest of the difference. We'll see how cooperative they turn out to be.

 

[RealNeil]

I really like to hedge my bet when I can. I look for an advantage wherever I can find it.
When it comes to fans, I like to use high quality, low noise, high flow cooling fans in place of whatever comes from the factory.

I've used Noctua fans to good effect before, (120mm & 140mm too) but I've had the best results with Cougar Vortec PWM fans. (set to run at top-speed because they're so quiet)
They have plenty of static pressure for radiators too.

You are confused, dude: OP's girlfriend has an 1920X in her PC.

Yeah, old people get confused, whoops!

 

[Valantar]

I really like to hedge my bet when I can. I look for an advantage wherever I can find it.
When it comes to fans, I like to use high quality, low noise, high flow cooling fans in place of whatever comes from the factory.

I've used Noctua fans to good effect before, (120mm & 140mm too) but I've had the best results with Cougar Vortec PWM fans. (set to run at top-speed because they're so quiet)
They have plenty of static pressure for radiators too.

Yeah, old people get confused, whoops!

I've generally only used Noctua for the past ... 10 years? Something like that. Got a couple of Be Quiet Silent Wings 3 and a Gentle Typhoon in my own PC currently as they were the best options at the time (as it was built before the NF-A12x25 was released, and previous Noctuas weren't that great on radiators). Cougar isn't sold in Norway, so that's not an option. If I had built the PC now, it would definitely be running NF-A12x25s all around. That's a custom loop, though, and as I said, it really shouldn't be necessary to replace the fans on a $140 AIO. If it was $80 or something, I'd be okay with it, but $140 is clearly "premium" segment pricing, and even if part of that is from the fancy pump and huge cold plate, they really could have gotten some better fans on there. I mean, they're supposed to be good - they're rated for a ridiculous 6.28 mm-H2O according to the spec sheet - but the noise level is definitely not the rated 28 dBA.

 

[Athlonite]

Noctua NH-U14S-TR4 is good for running at stock. Noctua NT-1 dont spread well on its own. You need to spread it.
Here is a video about TIM application on TR4:

And on HWinfo tdie is the real temperature of the cpu.

sorry but that guys an idiot he should have removed the cpu from the socket as all he was doing was getting tim from the edges between cpu and socket nothing was from the top of the cpu

 

[HTC]

I've generally only used Noctua for the past ... 10 years? Something like that. Got a couple of Be Quiet Silent Wings 3 and a Gentle Typhoon in my own PC currently as they were the best options at the time (as it was built before the NF-A12x25 was released, and previous Noctuas weren't that great on radiators). Cougar isn't sold in Norway, so that's not an option. If I had built the PC now, it would definitely be running NF-A12x25s all around. That's a custom loop, though, and as I said, it really shouldn't be necessary to replace the fans on a $140 AIO. If it was $80 or something, I'd be okay with it, but $140 is clearly "premium" segment pricing, and even if part of that is from the fancy pump and huge cold plate, they really could have gotten some better fans on there. I mean, they're supposed to be good - they're rated for a ridiculous 6.28 mm-H2O according to the spec sheet - but the noise level is definitely not the rated 28 dBA.

I've been using nothing but Noctua fans and coolers ever since i "discovered" the brand: i think it's been @ least 10 years.

Expensive for sure:



But the quality is there. So far, i've yet to have a single problem with any of their products: own 2 coolers (NH-C14 and NH-C12) and several fans besides those that come with the coolers. Add to that: they send you free of charge processor adapters, where possible, provided you can prove you have the product it's intended for.

 

[RealNeil]

I enter a lot of site contests and sometimes I get lucky. I won six 140mm Noctua fans and three 120mm Noctua fans from another site that I participate in.
I used them on a 6600K and later a 7700K build. All of them were connected directly to my PSU and running at full speed.
Here's a picture.

The system has since been decommissioned and the chassis is sitting on a shelf with all of the fans inside of it, waiting for another build.
__________________________________

Here is another build that uses all Cougar 140mm fans inside. It's in a Corsair Air 540 case. This is the i9-7900X build with 64GB of DDR4-3200MHz. RAM in it.
There is another mainboard in it now. (An ASUS ROG Rampage-VI Extreme) Notice that with the MSI board, there was only room for one side of the 280mm AIO to be configured in push/pull. The CPU has been delidded, so it runs a lot cooler than before. With the ASUS board installed, there are four 140mm fans on the radiator now.
All of the fans are directly connected to the PSU and running full speed. The Cougars are quieter than the Noctua fans.

 

[IceShroom]

sorry but that guys an idiot he should have removed the cpu from the socket as all he was doing was getting tim from the edges between cpu and socket nothing was from the top of the cpu

You are calling Kyle Bennett an idoit!!
Here he is just showing how you should apply TIM and as result he is not cleaning properly.

 

[Valantar]

I've been using nothing but Noctua fans and coolers ever since i "discovered" the brand: i think it's been @ least 10 years.

Expensive for sure:

View attachment 108595 View attachment 108596

But the quality is there. So far, i've yet to have a single problem with any of their products: own 2 coolers (NH-C14 and NH-C12) and several fans besides those that come with the coolers. Add to that: they send you free of charge processor adapters, where possible, provided you can prove you have the product it's intended for.

Yeah, they're not cheap, but quality is excellent. I have two of their 92mms in my HTPC/NAS that have been running 24/7 for the last 3+ years with zero issues, an old NF-P12 that served for 6-7 years in a previous build (and still works fine, I just don't have a use for it), and another NF-S12 that I have away along with some hardware I sold that had been in use for at least as long. This is my first experience with their CPU coolers, though.

I enter a lot of site contests and sometimes I get lucky. I won six 140mm Noctua fans and three 120mm Noctua fans from another site that I participate in.
I used them on a 6600K and later a 7700K build. All of them were connected directly to my PSU and running at full speed.
Here's a picture.
View attachment 108598

The system has since been decommissioned and the chassis is sitting on a shelf with all of the fans inside of it, waiting for another build.
__________________________________

Here is another build that uses all Cougar 140mm fans inside. It's in a Corsair Air 540 case. This is the i9-7900X build with 64GB of DDR4-3200MHz. RAM in it.
There is another mainboard in it now. (An ASUS ROG Rampage-VI Extreme) Notice that with the MSI board, there was only room for one side of the 280mm AIO to be configured in push/pull. The CPU has been delidded, so it runs a lot cooler than before. With the ASUS board installed, there are four 140mm fans on the radiator now.
All of the fans are directly connected to the PSU and running full speed. The Cougars are quieter than the Noctua fans.

View attachment 108599

Those look like NF-P12s, I think? I've been very happy with the one I have. Is there any benefit from running it in push-pull on a radiator that thin, though? Also, you must have a higher tolerance for noise than me, running six (wait. I count seven?) fans at full speed constantly. I run the fans in my build relatively slow to keep noise down, but of course they ramp up with temperature. My CPU can hit 65-ish under load, but the GPU never exceeds 50. Water temps don't exceed 45 degrees or so, which I think is fine. Also, I'm never going back to cases large enough to fit that many fans I'm 100% sold on SFF, though the H200i doesn't technically qualify (it's above 20l of volume), I'm definitely not going back to mid-tower or larger cases. Just don't need the space. Still, interesting to see that you're able to keep something that big relatively quiet

 

[RealNeil]

Still, interesting to see that you're able to keep something that big relatively quiet

I like larger cases just because of the increased numbers of fans. LOL!
In that build, they were all 140mm in size except for the one on the bottom. (140mm fans are quieter than 120mm fans)
Too bad you can't get Cougar fans where you live. They really are quieter than the Noctuas are.

 

[Valantar]

I like larger cases just because of the increased numbers of fans. LOL!
In that build, they were all 140mm in size except for the one on the bottom. (140mm fans are quieter than 120mm fans)
Too bad you can't get Cougar fans where you live. They really are quieter than the Noctuas are.

Are those noctuas 140mm? Wow, that case must be big enough to fool my eyes then

Back in the day, I had a CM Stacker 830 (black-and-green "Nvidia edition" despite housing first CF Radeon 4850s and then a HD 6950), which IIRC fit 7 120mm fans stock and probably 10 if I had made an effort (the front only had a single fan bracket, but plenty of room). I couldn't afford to stuff it full of Noctuas, but I had a few, yet the case was dusty, noisy and a major hassle. The Define R4 I replaced it with was better, but made it hard to accommodate the radiator for my Fury X (which was whiny anyhow, which is why I ended up putting a block on it and building a full custom loop). By that time, I had downsized from triple RAID5 HDDs and dual GPUs to a single GPU and a couple of SSDs. Don't see any reason to go back. I could never get a "lots of quiet fans in an open-air case" build to actually be quiet (even at low speeds, as the GPU and CPU fans would inevitably get noisy) which is why I went the more controlled route of custom water. If I want over-the-top cooling I can ramp up my fans (which still only make a pleasant 'whoosh'), but for now I'm happy with running them as slow as I can while maintaining decent temps. It's not a cheap solution, though, but I couldn't get cooling this good with this little noise in a case this small otherwise.

 

[RealNeil]

Are those noctuas 140mm? Wow, that case must be big enough to fool my eyes then

All but the one on the bottom are 140mm.
The chassis is an NZXT Source 530 Full tower. Here is TPU's review on it.
The airflow is pretty good in it and it's pretty solid too.
But I've outgrown it and I'm moving on to others now, so I'll probably give it away or donate it somewhere.

As for noise, I'm very hard of hearing, but when I have my hearing aids in, I can really hear airflow noise quite well.

I'm building in a Rosewill Cullinan MX Chassis today with eight RGB 120mm fans in it.
IDK what they'll sound like, but I'm hoping that I can tolerate their noise.
I know they'll be louder than the cougars that I usually use but I'm trying out an RGB build this time. This case looks sharp when it's lit-up.

EDIT: I just fired up the i7-8700K that I'm putting together in the Rosewill Chassis. I'm surprised at how quiet it is.

 

[micropage7]

noctua is good but i agree quality needs equal price

 

[Athlonite]

You are calling Kyle Bennett an idoit!!
Here he is just showing how you should apply TIM and as result he is not cleaning properly.

just look at the way he keeps assuming the TIM is all coming off of the top of the IHS when it clearly is not it's coming from all of that TIM between IHS and Retention bracket but does he open it and remove the CPU clean it and retention bracket separately nope then uses a micro fiber cloth and gets just as much TIM on that too oh yeah that's clean pffft not. So yes I'm calling Kyle Bennett an idiot because he is

 

[Valantar]

A bit of a self-necro here, but at least it's relevant.

I finally got around to ordering that second fan (they were out of stock in pretty much all Norwegian stores for a month or more, and I couldn't live with pairing a Chromax fan with a beige one, so I waited). I even did some quick-and-dirty before and after testing. I really wish I knew of some sweet data logging utility so I could get nice graphs like reviewers have, but at least I have OCCT's log images. There's a definite improvement: before, the CPU throttled down by a few hundred MHz, and didn't recover. This run was only around 10 minutes, but I know from experience that it would stay at those clocks (3.25-3.4-ish GHz).

Now, with the second fan, there are periods of very minor throttling, but they pass, and temperatures even drop back down to 65-66 degrees under full load for brief periods. I can't explain the fluctuation, but it seems like OCCT's power draw isn't as even as one would want it to be. Note the much longer run time. The first bout of minor throttling was at almost exactly the same time as in the previous run (weird - might this speak of some sort of pattern in terms of OCCT's workload?), but recovered quickly, and didn't dip as far.

Frankly, I'm a bit surprised that a second fan made a noticeable difference on a heatsink of this size, but I suppose it's big enough that relatively low-pressure fans like this can need some help.

IIRC someone also asked for pictures of the cooler in the case, so here are a couple. Frankly, this couldn't be a better fit. There's just enough clearance, yet no issues installing the cooler or fitting the side door back on. Have to love manufacturers that give out correct specs In the beginning I was worried that the fan mounting clips on the Noctua would touch the back of the GPU, but there's a few mm there too. And to my bafflement, it doesn't interfere with any RAM slots, regardless of fan mount height. Noctua must have been rather prescient when designing this cooler.

 

[eidairaman1]

A bit of a self-necro here, but at least it's relevant.

I finally got around to ordering that second fan (they were out of stock in pretty much all Norwegian stores for a month or more, and I couldn't live with pairing a Chromax fan with a beige one, so I waited). I even did some quick-and-dirty before and after testing. I really wish I knew of some sweet data logging utility so I could get nice graphs like reviewers have, but at least I have OCCT's log images. There's a definite improvement: before, the CPU throttled down by a few hundred MHz, and didn't recover. This run was only around 10 minutes, but I know from experience that it would stay at those clocks (3.25-3.4-ish GHz).

Now, with the second fan, there are periods of very minor throttling, but they pass, and temperatures even drop back down to 65-66 degrees under full load for brief periods. I can't explain the fluctuation, but it seems like OCCT's power draw isn't as even as one would want it to be. Note the much longer run time. The first bout of minor throttling was at almost exactly the same time as in the previous run (weird - might this speak of some sort of pattern in terms of OCCT's workload?), but recovered quickly, and didn't dip as far.

Frankly, I'm a bit surprised that a second fan made a noticeable difference on a heatsink of this size, but I suppose it's big enough that relatively low-pressure fans like this can need some help.

IIRC someone also asked for pictures of the cooler in the case, so here are a couple. Frankly, this couldn't be a better fit. There's just enough clearance, yet no issues installing the cooler or fitting the side door back on. Have to love manufacturers that give out correct specs In the beginning I was worried that the fan mounting clips on the Noctua would touch the back of the GPU, but there's a few mm there too. And to my bafflement, it doesn't interfere with any RAM slots, regardless of fan mount height. Noctua must have been rather prescient when designing this cooler.

Hmm i wonder if my cooler could run it. It has 2 fans and there isnt much difference in the 8350, I can actually take 1 fan off of mine.

 

[Valantar]

Hmm i wonder if my cooler could run it. It has 2 fans and there isnt much difference in the 8350, I can actually take 1 fan off of mine.

They look very similar except for the cold plate, with very similar dimensions (145x65mm vs 150x52mm - I didn't find fin counts), so if the Ashura had a cold plate suitable for TR4 I'm guessing it would perform pretty much the same - dependent on the fans, of course.