Spending £500+ on a cpu logically requires 20% of the cpu budget on a cooler; So around £100+ mark to be safe = Noctua NH-D15 Chromaxx (dual tower), two fan cooler; Arctic Freezer 2 280MM/360mm AIO; MSI Core liquid C280; EK AIO 360 D-RGB 360mm; etc. That being said, I didn't expect the Arctic freezer 34 e-sports cooler to do that well, GG to Arctic.
I got a combination of a well ventilated case, a Fractal Torrent, and a Noctua D15S for my i9-12900K and I've been running it hard in VR. I'm getting CPU 40C and GPU 60C. I've had a couple Noctua D15 cooler''s in other computers including an i9-9900K and they've done very well.
I don't mean to be rude, but not including NH-D15 in the test makes it pointless. It is the best air cooler and should always be used as a representation of the air cooling side. I really hope you make one more video just showing how the NH-D15 performs with the 12900K
I’ve been getting good results on a 12700k using a 120mm Thermalright assassin king 5 heat pipe. I switched out the fan with a Be quiet 120mm light wing 2500 rpm fan. I set the turbo limits in the bios to 190 watts, the cpu cap without over clocking, and it runs cinebench at 82C with a 70°F room temperature. Pretty good for a single tower cooler.
When I first installed my 12700k I was getting 230-240w and was hitting 96° in cinebench r23 in stock configuration on a Arctic liquid 360mm. I was using a rigged 212 evo backplate/standoff setup (which actually sits lower at 12mm instead of 13mm on the lga 1700 mount) I just figured it ran hot. Turns out the high wattage was being caused by insufficient cooling due to improper seating. Once I installed the lga 1700 mounting system (after 2 months of waiting for it to arrive) my temps are 70° max (10 minute run) and my wattage doesn’t break 200 with all power limits removed and running at sustained 4.8Ghz turbo.
Man, my i9 12900k is hitting 100 degrees instantly as soon as I start the benchmarch on Cinebench. I noticed that the power usage is passing 275w. I'm using a MSI Coreliquid 240R V2. Stock settings. What do you think might be happening?
would have been amazing video if you tested air cooling classics such as nhd15 or darkrock pro 4 on open bench and later in a case to see the real results. It was interesting, fun video but it could have so much more useful info for potential 12900k buyers.
Hi, I am quite confused why Noctua NH-D15 (either version) was not included, but I would be more interested in testing i7-12700k, after all, if you can afford i9-12900k, the cost of sufficient cooling should not be an issue, if yes, then you should go to i7 with better cooler, I think, it should give you better performance then.
@SharQ Well said! However, you don't actually need that boosting potential an aio would give in gaming with the 12900k, as it is actually cool during gaming. You can easily have a stable 5.1ghz in games. 12900k gets hotter in heavy multi core operations, but in gaming it shows impressively cool performance. D15 also has great headroom on the 129000k according to Noctua's compatibility chart.
It seems like a random mix of coolers with no real rhyme or reason. For both Noctua and Dark Rock, the coolers chosen are their weaker/slimmer designs rather than the "flagship" high performance air coolers. The likes of the Noctua D15 or the BeQuiet Dark Rock Pro (instead of the TF2) are actually designed to be high performance air coolers with enough performance to handle power hungry processors and/or serious overclocking. It would have made much more sense to test THAT class of coolers instead of random entry level ones with an i9.
Main thing is that the i9 doesn't really bring anything more for gaming over the i7, it only has 4 extra little cores(those help for heavy multithread load, 80% of games can't even use 8 cores properly) If you want you can OC the i7 to 5ghz and you will get same gaming performance as with the i9.
WOW! I own U14S (on threadripper), but since it looks so small and weak i ordered Freezer II 280 for my new 12900K instead. Maybe I didn't have to. Noctua is really killin it. Great and super useful video.
@@TheftTv oh my... thats insane! i will know in a week or so, when all of the components arrive. Although, even Freezer 240 is performing in benchmarks better than Noctua D15 for example, so theoretically it would have to be enough. We'll see
yeah just bought one for my 12700k build in a corsair 220t case so hopefully temps will be good enough. Im also gonna slap a extra fan on the back to help it out. What do u think
You should really check out the Scythe Fuma 2. It is $55, is only 155mm in height (so it can fit in all cases basically, unlike Noctua DH15), and it performs as good as Noctua DH15 (almost, like 1-2 degrees difference)
I just built a 12900k rig and paired with a cm hyper 212 air cooler, keeps it around 70c under extended load.. I’m honestly shocked how well it’s working
Huh, I'm trying to plan out my rig's components and my first plan was using an 12900F with a CM Hyper 212. This comment makes me feel somewhat better about that haha
wtf, is your hyper 212 boost in steroid? I have à 5800x with hyper212 and i run cinebench at 90 degrees and 135w output lol. But i have only one fan on it, a be quiet silent wing 3.
Exactly! Also, even if they perform well inside a case, the situation will be very different once you have your GPU blowing additional 300W of heat from below.
Conscientious testing done in a specific case, built with specific components along with how the build was finished for testing, suffers an increasing lose of credibility, as the comparing build deviates more.
as transistors heat up they leak more current, which is probably why you're seeing more power usage at temp limit. on the other hand, running chips cooler also helps with stability, which means that there's a bit more undervolting potential, resulting in even less power usage
my 12700K runs on the Noctua NH-D9L (single fan), that I took from my 6700K. With room temperature at 20° C, the cores run between 27°C and 33°C, with the center core 4 at 41°C, when running 4k120hz. When I switch to 8K60hz, temps go up to 31°C-35°C, with center core 4 hitting 49°C and 5 at 45°C. Much less than I expected and not even under-volted.
Good video, one thing you did not do was put the setup in a case, which will make a big difference. if you revisit this again do with an aio and have a temp sensor on the water. That would be a fun watch
according to " noctua compatibility centre " a nh-d9l air cooler ( only 11 cm high ! ) can cool a i9-12900k with max room for boost and overclocking . It is a lot smaller than other cooler that can do the job . Is this really possible ? Can you test it ?
as other people in the comments I have a couple of requests XD but doesn't include the NH-D15. There are two budget coolers that I've been using and I consider it will be interesting for you to consider: Thermalright Peerless Assassin 120 (starts in $36usd and go to $54usd with rgb) and Vetroo V5 ($35usd)
Remember that open test bench always will have higher temps than good ventilated PC Case - its normal. I have 3 push, 3 pull 120mm in case.. and on both CPU coolers what i had temps were much lower than in any test with same CPU. And yes it can be - without testing DRP4, Noctua 15, Deepcool AIII - im more than sure, that they will handle with that.
Temps of your system depends not just on the cooler u have, it also depends on the temperature of your country, here in Pakistan, summers are as hot as 50 degrees celsius and winters are as cold as 4-5 degrees, so cooling is hugely better in winter but in summer, no matter what you do, temps will always be high 😢
Erm, the arctic 34 duo was actually better than most of the others, yet you say you wouldn’t use it? Why? If you like, you can use it and pay another $100 to a friend or random person so you feel it’s worth using….
I went with watercooling. 12900K just running Asus auto OC @ 68% and seems to be peaking around 80°. It definitely jumps quickly though from its idle 30° to that 80°. These temps are overly high still IMO, definitely would investigate custom higher quality water blocks in the future, the corsair one I got could be better at transferring heat.
I’m ngl I give this props considering my acer nitro 5 laptop with a intel core 9th gen i5-9300H runs at almost 90° C sometimes on a standard cooler idk if it’s stock or not but it’s a ninth gen intel
My 12900K is on 5.1 all core (5.5 2 cores, 5.4 4 cores, 5.3 7 cores, ring & ecore at 4.1) on a NH-D15 for the moment, no problem at all with the cooling with Ambient temp at 18C (winter temp 😁) It is in an open air bench as of now & sitting in the "air pushing to gpu" position (the vrm heatsink of the strix A D4 doesnt allow it to sit on the normal position) with the standard LGA-115X bracket.
You manage to come up with the most interesting video ideas! I would probably use the largest Noctua that would fit in my case. Black fans a must, though.
The Be Quiet! Pure Rock Slim 2 - is the cooler I use on i3s and i5s.. its not that they really need an aftermarket solution(although you will get better temps) its so your build looks the coolerest!
The only air cooler I'm even considering purchasing for my i9 12900K would be the Noctua NH-D15 but sadly every place here in Canada is out of stock at the moment. For water coolers I've been looking at either Aorus Waterforce X 360 AIO or the NXZT Kraken Z73. MSI has the MEG CORELIQUID S360 which is LGA1700 compatible right out of the box. All these coolers are super pricey however and in short supply atleast here. If you have an i5 or i7 you can probably get away with a 280mm AIO but for the i9 I'm definitely gonna go with 360mm fans.
Lol air cooling an I9 12th gen? Are you trying to give your cpu an early death. Water is the only sensible way to go and don't get the kraken. Get the artic freezer ii aio. Its the best aio
You might wonder what is realistic. An open test bench is not what is giving you the right answer when you ask if you can cool it. People will have it in a case. That case might be under a desk. That desk might be in a room without airco in the summer. That room might be small like 3 x 3 meter (10 x 10 feet) and therefore the PC might even heat up the room. Since I don't expect people to work in rooms that are much hotter then 35 - 40 degree Celsius, that is roughly what should be tested. There is no point in testing a cooler in the winter on a test bench if the PC will also be used in the summer under a desk.
everybody told me it would be impossible to air cool a 9900kf when it released especially in a midi tower however its running fine for almost 2 years now with the same cheapest air cooler you showed in the video peaking at around 89 degrees
Speaks about 12900K + air cooling and does not test the NH-D15, aka the best air cooler on the market. I fail to understand why you wouldn't test this one. Someone packing a 12900K wouldn't cool it with a 25 bucks cooler.
I have an asus tuf z690-plus-wifi d4 with an i7 12700k and 32gb g.skill 4400 c19. After update bios to 0707 with the same watercooler my cpu dont pass 47 c. The memory and bios update change all my system. This cpu works well if memory is up to 4000mhz (ddr4). With my 3600mhz kit i reached easily 80 c. You can investigate this.
It's what you want to do is just a game, then I would say the $25 cooler would work perfectly. Believe it or not you could get the 12900K to run properly on that $25 cooler. You were just need to undervolt it.
I would like to see the 12900k on the fuma 2 revB. i have this on an 11900k and at 5ghz @ 1.22v all core pulling 170w it never goes above 70°C in 22°C ambient. The biggest hold back was the fans. I paired it in a nr200p with a Noctua NF-A12x15 and NF-A12x25 set to 1500rpm at max load. My guess is at 240w there would be a Δ of 71°C. Id love to see it for the science. maybe do a SFF dual tower air cooler shootout on the 12900k.
A bit unrelated but I am trying to keep a Ryzen 3900x cool with that BeQuiet! Pure Rock Slim 2 in a small case in which bigger air coolers aren't going to fit. Basically, what worked was turning mode off turbo mode and setting a custom power limit in bios, currently at 110W but could set it a bit higher. The whole reason for kneecapping it a bit and using this cooler over stock is that the stock cooler was so damn loud. But, stock performed better. It definitely can't handle this 12900k while it is pulling >200W xD
I had literally just ordered a Noctua NH-U12A before watching this, but all of the comments I saw in this video about the NH-D15 caused me to cancel my order and order the D15. Also saved me like $30, because the D15 is on sale now. I only ordered the U12A for the appearance, but if the D15 has superior performance, and I'll save a few bucks, I guess I can roll with that.
Unfortunately you didn’t show the best air cooler on the market which after watching a video clearly shows THE NH-D15 can handle the heat because max temp was 72c with a 10 minute loop on Cinabench.
i’m so glad i can reuse my NH U14S! I’m probably overthinking it but i’m not sure if the old “pea sized dot of thermal paste in the middle” tradition is still best here since these chips are now rectangular instead of a square. Anyone have any input on this?
@@weescamp when i ordered the LGA1700 adapter it came with a pamplet showing the best pattern for the new socket which was neat. I did go with it and I'm very happy with it.
Would love to see a sub 150mm shootout with that CPU, for those with height limitations...Thermalright Assassin King 120 SE vs Noctua NH-D12L vs Thermalright Silver Soul 135. Thanks!
Thank you for the video. Didn't expect such a bad result from dark rock tf. Wanted to buy this cooler in cooler master nr200 for 12600 or 12700 without overclocking. But it looks like he will not cope with them either.
The dark rock TF 2 is strange... it works differently from other cpu fans, it does not extract the airflow but absorbs it... many users install it wrong but from the same page they explain how it works... for this reason it also depends a lot on the chassis fans.
My 12900K cooled by a D15 was hitting 100 and throttling. In CB23. Switched XMP back to Auto and dropped temp to low 90's. Some undervolt with good effect.
having the down draft cooler on a flat test bench is pointless, its just blowing the hot air back at the cpu causing it to seem so bad. interesting to see how it would perform on a vertical test bench.
My experience with Noctua fans extends to a number of their NF-A9 PWM and NF-A14 PWM fans. However, their quiet operation in the still of the night, their non-offending operational sound(s) as they rev up to full speed, plus their wide dispersion of the air exhausted, leads some to conclude that these fans do not perform well, and cost too much. Yet, I have experienced Noctua fans that have disappointed me. As a case exhaust fan at the rear of the left side panel, in my mini-tower, the narrow NF-A9x14 PWM fan produced an unpleasant dronish sound starting at ~1200 RPM. Above 1700 - 1800 RPM, the drone-tone was intolerable. Tried the narrow NF-A9 HS-PWM Chromax fan. It produced an even stronger drone-tone starting at about 1100-1200 RPM. Above ~1600 RPM, the dronish tone was too uncomfortable to tolerate. Instead, I very carefully trimmed down the thickness of a NF- A9 PWM fan. This provided the 2nd exhaust fan an adequate clearance to inhale interior air for exhausting to the exterior. Since my very first build, I have felt it prudent to verify the contact patch between the processor IHS and the mating bottom of its cooler. I keep the cooler retain in contact with the CPU until the following day, making sure that the CPU hade been stressed the day before, in order to heat this contact area. Carefully lifting off the Noctua NH-U9S cooler, I did observe the Cap C contact patch now brought to public attention by many. Intel reports publicly, that It will void a previously assigned RMA, if the returned CPU, lacks clear manufacturing identification markings - unambiguous! Noctua has replied to an email, stating that removing the slight central hump, (0.002 - 0.003 inch, if I remember correctly), on the bottom surface will reduce the cooler’s effectiveness slightly. I first draw out a 1/4 inch of Noctua NT-H1 paste and discard. I then apply 6 - 8 small drops on the cleaned and secured CPU IHS. Placing a strip of Scotch Magic Tape over my clean index finger tip, I spread a VERY thin layer of NT-H1 paste over the entire CPU IHS, wiping any excess on a clean Bounty paper towel, or adding a bit more. Satisfied that the IHS is 100% wetted with a MINIMUM of paste, I repeat the process on the precise mating surface of the cooler bottom, but NO FURTHER OUT. I have found that the following careful mounting procedure, has created a more desirable contact patch between the CPU and its cooler in my PC. On the side making poor contact, I engage the screw under spring tension 1.5 - 2.5 turns. Then, one side to the other, back and forth, I tighten 0.5 turn at a time, until neither screw can be tightened any more. I also tighten the Intel arms until the hand tighten thumb nuts are JUST bottomed out.
I got I7-12700k and bios settings PL1=PL2=190 W and cooling NH-U12A and its 75C max. The test should also be done with Intel "stock settings" PL1=125W and PL2=240W (56sek). The PL1=PL2=240W is only if it can be cooled (yes Intel "allows" PL1=PL2=240W if cooling is ok).
Dear dude. Could you try noctoa c14s cryrig c1 and Raijintek Pallas both are 140 cooler, test them with stock fans, then phanteks 30 artic pmw and noctoa 2000rpm and 3000rpm pmw. Thanks in advance very much. Appreciate your work.
I was wondering if you had undervolted this 12900k on this test? Other sources say and almost all that even noctua NH-d15 can't save this from throtte, users on reddit say the same. Can you re-stress the 12900k on your noctua NH-d15 on cb r23 for like half an hour.. fast forward? There's not a single such video available in RU-vid. It'd be a massive help!!! Please do this
The bios has 3 Automatic options to throttle the watt. Stock cooler, air cooler and watercooler.. at air cooler option max Watt will be 240w. At watercooler it will be 4096w (which just means That there is no throttle)
You didn't run the AIO long enough to thermally saturate the liquid and reach equilibrium. I'd be very surprised if it didn't throttle after an hour or so.
You can also turn off the turbo boost function in the bios or under volt the CPU. Keeping the CPU at base clock speeds will reduce power draw and heat temperatures down. Another option is to set all cores at 4.2 GHz.
I’m quite interested to see this test with different all in one liquid coolers, as I’ve bought a 12900k myself paired with a lian li galahad 360mm with a lga 1700 mounting bracket for that cooler. I have similar clock speeds and wattage in this video and it heats up to 100c in a first single cinebench run in no time, it’s not my case cause it’s a phantek’s enthoo 719 with 9 fans. So I’m trying to figure out if this is a aio issue or something else possibly, cooling test info is pretty limited currently for i9 12th gen and this would be really helpful!
Wow I'm a bit surprised it gets to 100c with all of those fans! I wonder if using 140mm case fans or 140mm high speed exhaust fans would help? Maybe push pull configuration on the radiator?
I am curious how it would look with a cheap(ish) aio radiator with noctua fans(which you can buy separately from the dh-series coolers)-I have installed quite a few of them in various systems, and find they are often able to reduce the temp inside the system by up to 10 deg(C)(Compared to cheap-ish default 140mm fans)
I am both surpriwed and proud as a Freezer 324 eSports Duo owner, I guess i wont have to worry about cooler when upgrading from Ryzen 3600 to Ryzen 5800x or maybe even next gen AM4 ryzen.
Yeah I have a 5800x and I stay at 55c when gaming all core over clock to the max clocks with a hyper 212. Honestly Id keep my hyper 212 even if I went with a 5900x
Yeah your room temp is 21C, this is unrealistic, in hot summer months without AC a room can get as hot as 27-30C, there is no way ANY cooler would keep up - and most air coolers would result in passing Tjunction temps.
Don't forget the temps will be much hotter in the real world scenario. Just imagine what will happen in a case, with the graphics card blowing additional 300W of heat at the CPU. I thought NH-U12S was enough to cool my 10700K. And it was.. until I put the GPU under load as well.
All of these coolers should be just fine for gaming. My 9900k would throttle on this test, but I never see above 80C in games. If you are gaming, you should want to throttle in a test like this because it means you are maximizing your overclock. Games run about 20-30C cooler.
This shouldn't be such a challenge, Intel should have got their chips down to 1v or less with the commensurate drop in power by now. I'm still somewhat bemused that my 14nm 9700K needs more-or-less the same voltage as the 32nm 2600K it replaced. Got the shock of my life seeing the at-the-wall draw jump straight to 290w with an AVX small FFT Prime95 test @ 4.8 all-cores 1.3v (and the cores instantly into the 80's - stopped it pronto). They just haven't really made anything but incremental progress year on year for over a decade. Agree about idle consumption mentioned in the INtel/AMD comparison a couple of weeks ago though - the 9700K sits at 44w (at the wall) compared to 54w for the 2600K (I think they actually use similar power idling but the Asus Prime z390 MB is an improvement on the P8Z77-V) .