• Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!

Cooling Top Down Cooler Shootout?

Kmpkt

Innovation through Miniaturization
Original poster
KMPKT
Feb 1, 2016
3,382
5,936
I'm using Realtemp to measure temps. I've also tried HWINFO64 and the information was similar.
 
  • Like
Reactions: AleksandarK

Broxin

Cable-Tie Ninja
Jun 16, 2017
188
135
Realtemp is from 2012. Try http://www.alcpu.com/CoreTemp/ they are constantly pushing updates to this.
anyhow, the temps are the same with Realtemp and Coretemp

AIDA on this stresstest however displays 10°C lower Temps...

i tested it like ceski mentioned
AIDA64:


and had 80°C on an stock i7 6700k 4211,30MHz , fan curve "Silent"



p.s. i delidded my cpu and applied liquidmetal for better temps. Didnt had that much of an effect, maybe 4°C
 
Last edited:
  • Like
Reactions: AleksandarK

Kmpkt

Innovation through Miniaturization
Original poster
KMPKT
Feb 1, 2016
3,382
5,936
Yeah I have a set of results for "Stress CPU" only and they certainly stratify the coolers effectively. I might just keep those results for the final report as the result on the "Stress CPU + FPU" results correlate with them quite cleanly.
 
  • Like
Reactions: AleksandarK

Broxin

Cable-Tie Ninja
Jun 16, 2017
188
135
And how again did you exactly stress the cpu before, where it maxed our at 90degree so i can replicate it?
 
Last edited:
  • Like
Reactions: AleksandarK

Kmpkt

Innovation through Miniaturization
Original poster
KMPKT
Feb 1, 2016
3,382
5,936
I ran AIDA64 with both the Stress CPU and Stress FPU boxes checked. Is Stress CPU a reasonable metric of real world cooling? If so, then I'll just complete my data sets using both with the emphasis on that test.
 
  • Like
Reactions: AleksandarK

Broxin

Cable-Tie Ninja
Jun 16, 2017
188
135
can a comparsion even be made?
i mean, every mainboard has different fan profiles that can even be tweaked per mb manufacturer in silent, normal, max, full load, etc.
so there is no common denominator
 

Kmpkt

Innovation through Miniaturization
Original poster
KMPKT
Feb 1, 2016
3,382
5,936
Great advice. My testing board is in the mail as we speak. Unfortunately last time it came back from RMA the M.2 was smashed. In light of these settings, maybe it was for the best.
 

Runamok81

Runner of Moks
Jul 27, 2015
446
622
troywitthoeft.com
This still on hold, eh? The Skyreach thread has some folks testing the Titan TTC-NC25 paired with a Noctua nf-a12. 30mm cooler with 15mm fan. Wonder where that slots in?
 
Last edited:

Kmpkt

Innovation through Miniaturization
Original poster
KMPKT
Feb 1, 2016
3,382
5,936
I'm moving my entire heatsink collection on to @IntoxicatedPuma in February and he will be finishing the piece on behalf of smallformfactor.net

I honestly just haven't had the time as I've been working 50+ hours a week to pay for KMPKT Designs projects, have three kids and two other side businesses as well. Got a little too ambitious :(
 

Runamok81

Runner of Moks
Jul 27, 2015
446
622
troywitthoeft.com
I'm moving my entire heatsink collection on to @IntoxicatedPuma in February and he will be finishing the piece on behalf of smallformfactor.net

I honestly just haven't had the time as I've been working 50+ hours a week to pay for KMPKT Designs projects, have three kids and two other side businesses as well. Got a little too ambitious :(

Well understood. The top down cooler shootout will be in good hands with @IntoxicatedPuma. He's already off to a nice start.
Also, loving all the stuff that's been coming out of your recent design endeavors! Your unique DC-DC solutions are amazing.
 

Pat-Roner

SFF Lingo Aficionado
Feb 18, 2016
140
103
Kinda curious: anyone here ever considered making/owning a CPU cooler heater test setup? Something like what Overclockers and Anandtech does?

Better to just use a CPU tbh, then people have a reference point which they can compare to if they have the same gen or something with similar tdp. I know you can simulate the same tdp with a test rig, but people would still have a hard time referencing - i think
 

zovc

King of Cable Management
Jan 5, 2017
852
603
Better to just use a CPU tbh, then people have a reference point which they can compare to if they have the same gen or something with similar tdp. I know you can simulate the same tdp with a test rig, but people would still have a hard time referencing - i think

I can see where you're coming from. I think there's value in both. I think it's more "technically" valuable to scientifically control the heat against the heatsink's contact for the sake of properly understanding how well coolers work in a... empirical(?) sense? There's obviously a lot of high-level conclusions that can be drawn from that info.

On the other hand, it definitely is a lot more relatable to say "[this cooler] gave [these results] with [this CPU]," and expect the average consumer to be able to make sense out of it. Whether it's comparing to hardware they own, or because they're considering the configuration tested. I think the two approaches cater to different recipients of the data.
 

jØrd

S̳C̳S̳I̳ ̳f̳o̳r̳ ̳l̳i̳f̳e̳
sudocide.dev
SFFn Staff
Gold Supporter
LOSIAS
Jul 19, 2015
818
1,359
Kinda curious: anyone here ever considered making/owning a CPU cooler heater test setup? Something like what Overclockers and Anandtech does?

SFFn snapped Noctua using this exact setup to demo performance differences between some of their fans at Computex last year. It would be interesting to know weather this is something they do generally or just something they threw together for that specific demo. IIRC there have been publications in the past that chose to go w/ a heater for their reviews as well, not sure if its still a thing these days though. As @zovc said, there are advantages and disadvantages to both approaches.

Right side of the bench, the two cubes under the two heatsinks
 
  • Like
Reactions: zovc and rcodi