Question Has anyone seen a post about the best AIO Water CoolerIs for I9-9900k

[andenlee]

Google is no help, so i cant find anyone who has tested multiple AIO's and a custom Loop on a I9-9900k OC/Non-OC to see whats best for thermals and how to achieve good thermals while still busting a good OC.

Current Build:
Fractal Meshify S2
Kracken X72 (3 Pull fan setup) (changing to EVGA CLC 280 this weekend)
I9-9900k (Liquid metal between the IHS and the AIO plate)
36Gz G.Skill TridentZ RGB 3200 (changing to non RGB version, due to issues with the RGB controller, this weekend)
EVGA FTW3 RTX 2080TI (Air Cooled)
Samsung 970 Pro 1TB
Samsung 860 Pro 1TB
ASUS ROG MAXIMUS XI Formula

Current temps and clocks are, 53/53/53/52/52/52/51/51 (in order by core) 1.45v (adaptive) with 98°-105° under stress testing and 86° under normal usage (video games, rendering, ect)

 

[Jeepman2004]

Darkbreeze, no insults where thrown to you. I do know who you are, at least on the web wise sense and respect your work. More to the people saying the 8700k was thermally better. The i9's approach to soldering could be improved with a thinner application, but still faires better vs the 8700k thermal paste between the die and heat spreader. The 9900k is a tricky chip to work with, thermally speaking. I will admit, I have a good binned chip and the results of mine could not be replicated on my buddies system without stability issues.

Phaaze88, when my tweaks where made, thermal and voltage stability were fine. Surviving a Prime95 9 hour evaluation. When I get home, I'll post the BIOS settings I settled with and have running with my i9-9900k.

 

[Phaaze88]

Phaaze88, when my tweaks where made, thermal and voltage stability were fine. Surviving a Prime95 9 hour evaluation. When I get home, I'll post the BIOS settings I settled with and have running with my i9-9900k.

What settings for Prime95? Prime95 just covers thermal stability.
It is not as effective for testing voltage stability with the likes of Cinebench R20 or Aida64, as the workload is too steady/consistent.

 

[Darkbreeze]

Prime95 really isn't a good stability metric in any case, regardless of which of the torture test presets you choose. It IS a good metric for thermal compliance testing, but is less so for stability. Don't get me wrong, it IS a good "second or third opinion" stability test utility if you choose the Blend mode or a custom settings test especially if you tend to run scientific or math based applications and especially on projects that might run for long periods of time, but it's not as good as say Realbench or OCCT, H.264, few others.

 

[Notorious^]

The age old debate continues lol.

I will venture to say that my system is probably not stable on AVX stress test for a long period of time but my thoughts on this age old topic is this, stressing my system for xx hours is just not a real world scenario for me.

There is nothing i run or do that keeps my CPU maxed out for hours. I game, surf the interwebs and benchmark.
I do adjust my BIOS settings and GPU OC accordingly but there is nothing i do with my PC that it can't handle at 5.0,5.1 & 5.2GHz. I can get a post at 5.3GHz but as soon as i put any kind of load i BSOD with the quickness and i don't feel comfortable throwing any more voltage at it.

I will admit the 280 RAD i had was not enough to cool with my OC's i was running during bench testing and swapped it our for a 360 RAD which helped, did a delid and switched to a 011 Dynamic case with 10 fans in this sucker. On my last Cinebench R20 @ 5.2GHz my max temp was 72C @ 1.29V with all fans running at 2400 RPM (Corsair ML 120's)

So i don't think its crazy to say that a user should plan & design to cool a system based on what they're using it for. Just my $0.02 cents.

 

[Darkbreeze]

It's not about the "hours", in most cases it is about specific instructions or specific conditions, which could be reached within seconds in a given application but might not be a part of a stability test utility until further into the process.

Furthermore, it's not necessarily about what YOU do, but in many cases is about what WINDOWS does, because it does it's own thing completely divorced from what you as the user may think is/are the only priorities for the system. So yes, it is crazy to say that "I'm just a gamer" or "I only browse the web" as an excuse for not ensuring stability. It's like saying "I'm only driving three blocks to the grocery store" so that tire that is 30lbs low is no big deal because I won't be driving at high speed. Well, yes, it IS a big deal, because you will probably do that repeatedly and eventually wear the tire bald in which case you will also probably end up having a blow out and either driving off the side of the road or into oncoming traffic when it does.

The cumulative effect of system instability is not much different than that in the end.

If you are going to overclock your CPU or memory, DO THE FRACKING WORK necessary to demonstrate that it is stable, or else don't overclock.

Now, if you have a system that has nothing but the OS and one game installed and that is all you ever do and you don't care if your saves become corrupted over time, AND don't mind frequently reinstalling Windows (In the case of memory instability), by all means, run your system on the frayed edge of sanity all you like. Or do it even if that's not your prerequisite, after all, it's your system and nobody can tell you what to do, right?