[GN] Why Most Cooler Tests Are Flawed - Page 2 - Overclock.net - An Overclocking Community
Forum Jump: 

[GN] Why Most Cooler Tests Are Flawed

Reply
 
Thread Tools
post #11 of 21 (permalink) Old 03-07-2020, 06:23 AM
New to Overclock.net
 
DNMock's Avatar
 
Join Date: Jul 2014
Location: Dallas
Posts: 3,708
Rep: 174 (Unique: 128)
That is a really long write up to say "Use common sense, and don't be lazy".


DNMock is offline  
Sponsored Links
Advertisement
 
post #12 of 21 (permalink) Old 03-07-2020, 06:55 AM
Tetrapyloctomist
 
Aenra's Avatar
 
Join Date: Feb 2017
Posts: 1,675
Rep: 44 (Unique: 27)
Two points of concern (though efforts such as this one are of course highly appreciated):

- They mention disregarding highest/peak single core temp as it can skew the data.
While accurate, means and averages and logisti-like accountant reminding "portrayals" -of what we think is best- always annoyed me. Because by actual fact, they hide what's of essence. In this instance, what your highest core temp will be, which i believe is the number one concern, is it not. I don't care if "on average" x or y cooler is better, i care specifically about said spikes and which cooler can handle those; said spikes being what's stopping me from higher OCing, lower DBs, or whatever else is a personal factor in my system.

- We've been through the non-CPU, non-mobo controlled external heater box method before. We also discarded it, over a decade ago, for some very valid reasons that aren't going anywhere either.
It has its own issues that it brings to the table, misrepresenting the actual data, and unless they have a fab willing to fabricate faux sub-10nm pieces of metal for them? Each of them individually powered? All soldered to a fixed dimensions substrate? I couldn't care less.
With bold font because judging by the comments above, we have a lot of people not knowing but happily commenting here nonetheless, lol, what a surprise: thermal density.
If their heater boxes accommodate for that and in a manner specified -read: shown publicly- so we can be convinced they actually do? Then and only then will i be impressed. And i really would be.

Until then, pink meme boxes reviews, teardowns and easy, cheap Walmart bashing... whatever.

Pride, honour and purity.

Last edited by Aenra; 03-07-2020 at 07:02 AM.
Aenra is offline  
post #13 of 21 (permalink) Old 03-07-2020, 07:47 AM
New to Overclock.net
 
DNMock's Avatar
 
Join Date: Jul 2014
Location: Dallas
Posts: 3,708
Rep: 174 (Unique: 128)
Quote: Originally Posted by Aenra View Post
Two points of concern (though efforts such as this one are of course highly appreciated):

- They mention disregarding highest/peak single core temp as it can skew the data.
While accurate, means and averages and logisti-like accountant reminding "portrayals" -of what we think is best- always annoyed me. Because by actual fact, they hide what's of essence. In this instance, what your highest core temp will be, which i believe is the number one concern, is it not. I don't care if "on average" x or y cooler is better, i care specifically about said spikes and which cooler can handle those; said spikes being what's stopping me from higher OCing, lower DBs, or whatever else is a personal factor in my system.

- We've been through the non-CPU, non-mobo controlled external heater box method before. We also discarded it, over a decade ago, for some very valid reasons that aren't going anywhere either.
It has its own issues that it brings to the table, misrepresenting the actual data, and unless they have a fab willing to fabricate faux sub-10nm pieces of metal for them? Each of them individually powered? All soldered to a fixed dimensions substrate? I couldn't care less.
With bold font because judging by the comments above, we have a lot of people not knowing but happily commenting here nonetheless, lol, what a surprise: thermal density.
If their heater boxes accommodate for that and in a manner specified -read: shown publicly- so we can be convinced they actually do? Then and only then will i be impressed. And i really would be.

Until then, pink meme boxes reviews, teardowns and easy, cheap Walmart bashing... whatever.
You seem to be someone in the know on this stuff. I've always felt like Stren at Xtremerigs was one of the more thorough reviewers on water cooling. Is that a fair statement?

edit: Wow, guess I am falling out of the loop, just checked in at Xtremerigs and they haven't updated their site in a year now. Anyone got suggestions on where to go for watercooling reviews?



Last edited by DNMock; 03-07-2020 at 07:56 AM.
DNMock is offline  
Sponsored Links
Advertisement
 
post #14 of 21 (permalink) Old 03-07-2020, 08:22 AM
Never Finished
 
airisom2's Avatar
 
Join Date: Aug 2011
Location: Within the Milky Way
Posts: 2,137
Rep: 273 (Unique: 210)
Pretty sure they just designed the TTV to validate their testing methodology, like how they used a thermal chamber to validate their case thermal testing methodology. They are using actual processors to benchmark these coolers.

1680 V2 4.5GHz 1.33v | G.Skill Trident X 4X4GB 2400MHz | R4BE | Titan Xp + Morpheus II
Wasabi Mango UHD550 | Rosewill Rise | Thermalright LGM RT
Logitech G600 | G440 | Tt Meka G-Unit Cherry MX Black+Double O-Rings
JBL LSR308s | Temblor T10 | Custom headphones | TEAC-UD501+Passive Preamp | Emotiva BasX A-100
OCZ Vertex 4 256GB | 51TB DAS | SS Prime Ultra Titanium 1KW
airisom2 is offline  
post #15 of 21 (permalink) Old 03-07-2020, 08:56 AM
u broke 666 rep :(
 
skupples's Avatar
 
Join Date: Apr 2012
Location: Fort Lauderdale
Posts: 23,499
Rep: 681 (Unique: 361)
people confuse case temp with ambient temp all the time.

here's how i've always had it, thanks to all the 2 prongs aquacomputer gives you.

1 on radiator in
1 on radiator out
1 for case temp
1 for genuine ambient (outside of case temp)
2x water sensors.

R.I.P. Zawarudo, may you OC angels' wings in heaven.
If something appears too good to be true, it probably is.
Best R0ach Quote of all time : TLDR: Haswell might be the last legit gaming platform unless mice get their own non-USB interface on some newer architecture.
KAIHL Switches < CHERRY Switches, 10/10, blind taste test @ work.

Last edited by skupples; 03-07-2020 at 09:00 AM.
skupples is offline  
post #16 of 21 (permalink) Old 03-08-2020, 10:08 AM
AMD Acolyte
 
qwertymac93's Avatar
 
Join Date: Sep 2009
Location: just here
Posts: 1,477
Rep: 76 (Unique: 70)
Frostytech has been using a somewhat similar testing methodology for many years now. I'm interested to see how the two sites compare in their data output going forward.

http://www.frostytech.com/testmethod_mk2.cfm

"We should make it clear that NVIDIA is a good company, and we respect their attempted transition into being a smartphone company." - Roy Taylor, AMD

BF3 STATS/BF4 STATS
z87
(19 items)
Media Rig
(18 items)
CPU
Core i5 4670k - [email protected]
Motherboard
MSI z87m GAMING
GPU
MSI R9 280 GAMING - 1.15ghz core/1.5ghz mem
RAM
A-DATA 2x4GB DDR3
RAM
PNY 2x4GB DDR3
Hard Drive
Crucial M500 240GB
Hard Drive
Seagate 3TB
Power Supply
Antec TruePower TP-550
Cooling
Xigmatek Gaia + Cougar VortexPWM
Case
Silverstone PS07
Operating System
Windows 10 Pro
Monitor
Viewsonic XG2401
Keyboard
CM Storm QuickFire Rapid
Mouse
Logitech G402
Mouse
Reflex Lab Pro 32 Heavy
Audio
Beyerdynamic DT 770 PRO-80
Audio
Yamaha RX-V663
Audio
Micca MB42x Bookshelf Speakers
Audio
SMSL Q5
CPU
AMD Phenom II 720
Motherboard
ASRock 870 Extreme 3
GPU
HIS 5750 (840core/1200mem)
RAM
2*2GB A-Data DDR3 1600
Hard Drive
Seagate Barracuda ST1000DM003
Hard Drive
Seagate Barracuda ST3000DM001
Optical Drive
Lite-On Blue Ray WH14NS40
Power Supply
Corsair CX-430
Cooling
Arctic Cooling Freezer 64 PRO
Case
NZXT Beta EVO
Operating System
Windows 10 pro 64-bit
Monitor
Vizio M43-c1
Keyboard
Logitech K400
Mouse
Gigabyte M7700
Audio
Sony STR-DN1070
Audio
Definitive Technology BP8 Tower Speakers
Audio
Definitive Technology 2002 Series Center Speaker
Audio
Definitive Technology ProMonitor 1000 Bookshelf Speakers
▲ hide details ▲
qwertymac93 is offline  
post #17 of 21 (permalink) Old 03-08-2020, 05:24 PM
Meep
 
Gilles3000's Avatar
 
Join Date: Feb 2013
Location: Belgium
Posts: 5,414
Rep: 327 (Unique: 234)
Quote: Originally Posted by Aenra View Post
Two points of concern (though efforts such as this one are of course highly appreciated):

- They mention disregarding highest/peak single core temp as it can skew the data.
While accurate, means and averages and logisti-like accountant reminding "portrayals" -of what we think is best- always annoyed me. Because by actual fact, they hide what's of essence. In this instance, what your highest core temp will be, which i believe is the number one concern, is it not. I don't care if "on average" x or y cooler is better, i care specifically about said spikes and which cooler can handle those; said spikes being what's stopping me from higher OCing, lower DBs, or whatever else is a personal factor in my system.

- We've been through the non-CPU, non-mobo controlled external heater box method before. We also discarded it, over a decade ago, for some very valid reasons that aren't going anywhere either.
It has its own issues that it brings to the table, misrepresenting the actual data, and unless they have a fab willing to fabricate faux sub-10nm pieces of metal for them? Each of them individually powered? All soldered to a fixed dimensions substrate? I couldn't care less.
With bold font because judging by the comments above, we have a lot of people not knowing but happily commenting here nonetheless, lol, what a surprise: thermal density.
If their heater boxes accommodate for that and in a manner specified -read: shown publicly- so we can be convinced they actually do? Then and only then will i be impressed. And i really would be.

Until then, pink meme boxes reviews, teardowns and easy, cheap Walmart bashing... whatever.
As far as I know, excessive peaks on a single core most of the time don't relate to the effectiveness of the cooler at all, but tend to be caused by imperfect factory TIM on the die or just a less than great die bin that's still within tolerances.

We'll see, but it looks like they're at least putting in the effort to get a consistent reliable setup going, which is more than you can say for the vast majority of reviews out there that don't even properly account for ambient and case temps, nor do noise normalized testing.

Cooler reviews are such a ****show atm, and the few people that do seem to do it properly seem to either be strapped for time, money or exposure. If GN can get a test out there that's even just mostly accurate that would be a huge improvement compared to what we have available now.
Gilles3000 is online now  
post #18 of 21 (permalink) Old 03-08-2020, 08:09 PM
u broke 666 rep :(
 
skupples's Avatar
 
Join Date: Apr 2012
Location: Fort Lauderdale
Posts: 23,499
Rep: 681 (Unique: 361)
Quote: Originally Posted by qwertymac93 View Post
Frostytech has been using a somewhat similar testing methodology for many years now. I'm interested to see how the two sites compare in their data output going forward.

http://www.frostytech.com/testmethod_mk2.cfm
i've gotta look at this more... but...

so instead of a chip, they're using a peltier to create set amounts of heat? oh i see. for total cooling potential of block, no other factors included (people care about other factors)

i feel like there are people around here that would complain about that due to "chiplet placement" and fin coverage, etc.

either way, I get the idea and methodology... interested to see them scale it up to 400-500w.

R.I.P. Zawarudo, may you OC angels' wings in heaven.
If something appears too good to be true, it probably is.
Best R0ach Quote of all time : TLDR: Haswell might be the last legit gaming platform unless mice get their own non-USB interface on some newer architecture.
KAIHL Switches < CHERRY Switches, 10/10, blind taste test @ work.
skupples is offline  
post #19 of 21 (permalink) Old 03-09-2020, 12:05 PM
AMD Acolyte
 
qwertymac93's Avatar
 
Join Date: Sep 2009
Location: just here
Posts: 1,477
Rep: 76 (Unique: 70)
Once you have an IHS and make sure you have good contact with it, chiplet placement doesn't really matter. "Direct touch" heatsinks might be affected a measurable amount depending on position of the heatpipes but coolers with copper baseplates have so much thermal mass over the whole thing I don't think chiplet placement would matter much.

Frostytech's methodology hasn't been updated for years, it might make sense for them to debut a "3.0" for epyc/threadripper platforms with output in the 250-300w range. The TR4 socket is so large and the chiplets so spread out it might make sense to penalize coolers that don't make full contact with the IHS somehow. At least, those that wouldn't cover all of the chiplets underneath anyway.

They don't use Peltier devices, BTW. The heaters are pure resistive loads, which are much more efficient.

"We should make it clear that NVIDIA is a good company, and we respect their attempted transition into being a smartphone company." - Roy Taylor, AMD

BF3 STATS/BF4 STATS
z87
(19 items)
Media Rig
(18 items)
CPU
Core i5 4670k - [email protected]
Motherboard
MSI z87m GAMING
GPU
MSI R9 280 GAMING - 1.15ghz core/1.5ghz mem
RAM
A-DATA 2x4GB DDR3
RAM
PNY 2x4GB DDR3
Hard Drive
Crucial M500 240GB
Hard Drive
Seagate 3TB
Power Supply
Antec TruePower TP-550
Cooling
Xigmatek Gaia + Cougar VortexPWM
Case
Silverstone PS07
Operating System
Windows 10 Pro
Monitor
Viewsonic XG2401
Keyboard
CM Storm QuickFire Rapid
Mouse
Logitech G402
Mouse
Reflex Lab Pro 32 Heavy
Audio
Beyerdynamic DT 770 PRO-80
Audio
Yamaha RX-V663
Audio
Micca MB42x Bookshelf Speakers
Audio
SMSL Q5
CPU
AMD Phenom II 720
Motherboard
ASRock 870 Extreme 3
GPU
HIS 5750 (840core/1200mem)
RAM
2*2GB A-Data DDR3 1600
Hard Drive
Seagate Barracuda ST1000DM003
Hard Drive
Seagate Barracuda ST3000DM001
Optical Drive
Lite-On Blue Ray WH14NS40
Power Supply
Corsair CX-430
Cooling
Arctic Cooling Freezer 64 PRO
Case
NZXT Beta EVO
Operating System
Windows 10 pro 64-bit
Monitor
Vizio M43-c1
Keyboard
Logitech K400
Mouse
Gigabyte M7700
Audio
Sony STR-DN1070
Audio
Definitive Technology BP8 Tower Speakers
Audio
Definitive Technology 2002 Series Center Speaker
Audio
Definitive Technology ProMonitor 1000 Bookshelf Speakers
▲ hide details ▲

Last edited by qwertymac93; 03-09-2020 at 12:23 PM.
qwertymac93 is offline  
post #20 of 21 (permalink) Old 03-09-2020, 05:48 PM
u broke 666 rep :(
 
skupples's Avatar
 
Join Date: Apr 2012
Location: Fort Lauderdale
Posts: 23,499
Rep: 681 (Unique: 361)
thanks for the clarification.

R.I.P. Zawarudo, may you OC angels' wings in heaven.
If something appears too good to be true, it probably is.
Best R0ach Quote of all time : TLDR: Haswell might be the last legit gaming platform unless mice get their own non-USB interface on some newer architecture.
KAIHL Switches < CHERRY Switches, 10/10, blind taste test @ work.
skupples is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off