OC3D's ryzen cooler testing - Overclock.net - An Overclocking Community

Forum Jump: 

OC3D's ryzen cooler testing

Reply
 
Thread Tools
post #1 of 39 (permalink) Old 04-25-2017, 09:33 PM - Thread Starter
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 10,046
Rep: 815 (Unique: 548)
https://www.overclock3d.net/reviews/cases_cooling/amd_ryzen_5_7_cpu_cooler_round_up/5
Quote:
When overclocking our R5 1600X we can see that the affordable Cryorig H7 was simply unable to keep up with the heat output of our R5 1600X at 4GHz. It seems that 1.4V core voltages are simply too much for small air coolers, with lower voltages and/or clock speeds being required for an overclocked 1600X to run on lower-end air coolers.


It's interesting in that a 140W TDP rated Cryorig H7 isn't enough for the 6 cores @ 4GHz.

Quote:
For our overclocked CPU testing we ran our Ryzen 7 1800X at 4GHz with a core voltage of 1.4V and a SoC voltage of 1.1V. We ran the CPU with our memory kit clocked at 3200MHz and ran the OCCT to simulate a worst case scenario CPU/thermal load.

In the below graph we can see that the Cryorig H7 has been taken off the graph, with the Noctua NH-D15S failing to keep temperatures under 70 degrees unless the fans were run at their highest speeds. It is clear then that heavy overclocks on our Ryzen 7 1800X will require an AIO liquid cooler to stay under 90 degrees without overly loud fans.

We can see that by using larger liquid coolers that we can decrease the load temperature of our 4GHz R7 1800X to 60 degrees or below, with the largest 280mm liquid coolers like the Corsair H115i and the NZXT X62 providing the best performance.


----

IMHO it would have been more interesting to see the following:

Deepcool Gammax S40 --- <$30
Thermaltake Contac Silent 12 (150W TDP rating) --- < $30 , high exposure on Amazon
Cooler Master 212 Evo w/ AM4 bracket --- $30ish , popular + Microcenter bundles bracket (supposedly 180W TDP rating , likely with 2K rpm fan speed)
Be Quiet Pure Rock (150W TDP rating) --- $30ish

Scythe Ninja 4 w/ AM4 bracket
Scythe Mugen 5 , Mugen Max , Mugen 4 , w/ AM4 bracket
Scythe Fuma w/ AM4 bracket
Cryorig H5 w/ AM4 bracket (rated for 160W TDP)

Be Quiet Dark Rock 3 w/ AM4 bracket (rated for 190W TDP)
Noctua NH-U14S SE AM4 (220W capable per http://noctua.at/en/tdp-guide)
Thermalright true Spirit 140 Direct , built after Jan 2017 (rated for 200W per Thermalright)
Thermalright Macho Direct , built after Jan 2017 (rated for 200W per Thermalright)
Thermalright Macho 120 Rev A w/ AM4 bracket (rated for 200W TDP per Thermalright)
Thermalright Macho Rev B w/ AM4 bracket (rated for 240W per Thermalright)
Thermalright TRUE Spirit 140 Power w/ AM4 bracket (rated for 360W)

alphacool Eisbaer with AM4 bracket / Be Quiet silent Loop <--- for Ryzen 7
Swiftech H220 X2 w/ AM4 bracket <--- for Ryzen 7
EK Predator 240 / 280 / 360 w/ AM4 bracket <--- for Ryzen 7


Of course there's a immense difference between a CPU at 1.4V versus one at 1.25ish V for 3.8 or 3.9GHz , since power goes up proportionally with square of voltage but linearly with frequency

----

Per Anandtech's findings on http://www.anandtech.com/show/11244/the-amd-ryzen-5-1600x-vs-core-i5-review-twelve-threads-vs-four/2 , dropping to 3.7GHz requires less than half the power of 4 GHz on the Ryzen 7 1800X / Ryzen 5 1600X

► Recommended GPU Projects: [email protected] , [email protected] (FP64) (AMD moreso) ► Other notable GPU projects: [email protected] (Nvidia), GPUGrid (Nvidia) ► Project list


AlphaC is offline  
Sponsored Links
Advertisement
 
post #2 of 39 (permalink) Old 04-26-2017, 03:20 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 24,854
Rep: 1632 (Unique: 916)
Interesting information. Thanks for posting.

But it is not a cooler to cooler comparison .. it is instead a comparison of how different coolers perform in a given system.

I started seeing red lights and and hearing alarm bells as soon as I saw OC3D results showing CLCs having 5c cooler temps than D15S at stock settings in your posted graphs. Quick look at review criteria shows their test system is in a Air 240 case in 'a temperature controlled room'. I'll assume this mean they are using room ambient temperature (20c) and not the air temp going into cooler during testing.

Their testing at the very best is only showing us how well their system (in Air 240 case) performs with different coolers. I don't have a system identical to theirs so their test results means nothing to me. I'm more interested in how coolers perform against each other .. as in each receiving the same temperature air going into cooler .. or at the very least using the air temp going into cooler as baseline for delta temps insread of the air temperature of room test system is in .. which obviously is not the same as cooler intake air temp. mad.gif

IMHO scratch that .. It is a fact that testing of coolers inside a system using air temperature of 'a temperature controlled room' for baseline air temperature gives us no idea how coolers are actually performing against each other. To know how the cooler perform we have to have testing based on cooler intake air temp. We need this to rule out the effects of case and system have on the airflow temp going into cooler. Therefore we have no idea how well each cooler performed, only how each cooler performed in their specific test system .. so unless you have an identical system OC3d test results are virtually worthless.

Using a case with 2x 140mm intake and 1x 140mm exhaust fans for air coolers versus 2x 140mm intake and 1x 140mm exhaust fans then adding the CLC fans to other vents as intake/exhaust changes case airflow dramatically.

Each air coolers' fan and airflow can also dramatically change the case airflow .. and any change in case airflow almost always changes the cooler intake air temperature

If we change the cooler intake airflow 3c, the CPU temp will be 3c warmer .. 6c warmer air into cooler results in 6c warmer CPU temp. While this is difference is not always exactly a 1:1 ratio, it is very close.

Custom loops air temp changes are not 1:1 .. it can be 5c change in air temp to 1c change in CPU temp .. all depends on components and water temp. I do not know how much closer to the 1:1 ratio CLCs are .. not even enough to guess.

But the changes in case airflow in air 240 with different air coolers is going to be noticeable .. I have seen 4-7c differences in cooler intake air temp depending on case airflow setup and what air cooler is being used. I'm sure their case testing of CLCs versus air cooler is at least as much more.

Last I knew the EK Predator is no longer being sold.
Mid-range coolers like H5, TRUE Spirit Direct, H7 (wish the new H7 Quad w/ 4x 6mm heatpipes was out), Macho Direct, etc do a very good job and are only a few degrees below top tier coolers lke Dark Rock Pro3, NH-D15S, R1, Silver Arrow IB-E, PH-TC14PE, TRUE Spirit 140 Power, NH-U14S, Archon IB-E X2, etc.
I would like to see testing of all the above based on airflow temp into cooler ..
adding Alphacool, be quiet! Silent Loop, Swiftech X2, etc all with same size radiator too. I don't care about CLCs and their test results. Maybe I'm too picky, but I want an AIO with a fill port, proper fittings, copper radiator .. a system I can maintain with components I can replace if needed. tongue.gif

Phanteks Owners Club Ways to Better Cooling
i7 980 @ 3.55GHz =PH-TC14PE w/2x TY-143 fans =Crucial Ballistix 3x4GB =GA-X58A-UD5 =ASUS GTX580 DirectCU II =Enermax Modu84+ =Define R2 w/3x TY-140 case intake fans; all PWM controlled by CPU fan socketPhanteks Enthoo Primo MoBo Rampage III Extreme CPUi7 980X
@ 4.0GHz =R1 Ultimate w/2x TY-143 fans =Dominator GT 6x2GB =Rampage III Extreme =ASUS GTX580 DirectCU II =TX850 =Enthoo Primo w/ custom castor base
doyll is offline  
post #3 of 39 (permalink) Old 04-26-2017, 08:59 AM - Thread Starter
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 10,046
Rep: 815 (Unique: 548)
The reason why I posted it is because it suggests a Cryorig H7 with 140W TDP isn't enough to cool a Ryzen 5 in a ventilated case. Many people had conjecture that 1.4V on Ryzen 7 is too much for midrange air cooling.

I suppose this is why Ryzen 7 review kits were sent out with the NH-U12S with industrial 2000RPM fans instead of the NH-U12S SE AM4.

If you look at the Anandtech page, cores at 4.1GHz (XFR) use ~ 22W , cores at 4GHz use ~21W while dropping down to 3.7GHz goes down to about 13W-14W. The memory controller and other non-core power appears to be up to 20W or so. This assumes the automatic voltage that Ryzen requests.
i.e. for 8 cores at 4.1GHz, we would expect 8 x 22W + 20W = 196W
8 cores at 4.0GHz : 8 x 21W + 20W = 188W
8 cores at 3.7GHz : 8 x 14W + 20W = 132W
6 cores at 4.1GHz : 6 x 22W + 20W = 152W
6 cores at 4.0GHz : 6 x 21W + 20W = 146W <--- Cryorig H7 TDP rating exceeded
6 cores at 3.7GHz : 6 x 14W + 20W = 104W

► Recommended GPU Projects: [email protected] , [email protected] (FP64) (AMD moreso) ► Other notable GPU projects: [email protected] (Nvidia), GPUGrid (Nvidia) ► Project list


AlphaC is offline  
Sponsored Links
Advertisement
 
post #4 of 39 (permalink) Old 04-26-2017, 09:26 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 24,854
Rep: 1632 (Unique: 916)
Quote:
Originally Posted by AlphaC View Post

The reason why I posted it is because it suggests a Cryorig H7 with 140W TDP isn't enough to cool a Ryzen 5 in a ventilated case. Many people had conjecture that 1.4V on Ryzen 7 is too much for midrange air cooling.

I suppose this is why Ryzen 7 review kits were sent out with the NH-U12S with industrial 2000RPM fans instead of the NH-U12S SE AM4.
While would not put an H7 on a Ryzen 7, it does not make their testing valid. The point is their testing proves nothing because we have no idea how 'well ventilated' the case is.

They give no data to validate the ventilation of the case except to say it has 3x 140mm fans in it .. we have no idea what the air temp going into cooler is, therefore we have no idea how good or bad the case airflow is.

While I think the NH-U12S with a 2000rpm fan is probably capable of cooling the Ryzen 7, why Ryzen review kits were sent with NH-U12S & IPPC 2000rpm fans could be any of a number of possible reasons. Maybe Noctua wanted to unload a bunch of NH-U12S coolers and an equal number of IPPC 2000rpm fans.

Here is a review showing NH-U1S being 2c warmer than H7 on an Intel Core i7-4770K @ 4.2GHz

https://www.techpowerup.com/reviews/CRYORIG/H7_Universal/6.html

But again, it is a test done in a case based on room air temperatures, not the actual temperature of air going into cooler.

Point is neither your or my review links and data prove anything about cooler against cooler performance.because neither review is testing coolers on their own .. but testing how coolers perform in a specific system.

Phanteks Owners Club Ways to Better Cooling
i7 980 @ 3.55GHz =PH-TC14PE w/2x TY-143 fans =Crucial Ballistix 3x4GB =GA-X58A-UD5 =ASUS GTX580 DirectCU II =Enermax Modu84+ =Define R2 w/3x TY-140 case intake fans; all PWM controlled by CPU fan socketPhanteks Enthoo Primo MoBo Rampage III Extreme CPUi7 980X
@ 4.0GHz =R1 Ultimate w/2x TY-143 fans =Dominator GT 6x2GB =Rampage III Extreme =ASUS GTX580 DirectCU II =TX850 =Enthoo Primo w/ custom castor base
doyll is offline  
post #5 of 39 (permalink) Old 04-26-2017, 09:41 AM - Thread Starter
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 10,046
Rep: 815 (Unique: 548)
Well as far as the TDP goes, I've seen many sources (i.e. screenshots of hwinfo64 and buildzoid's testing) stating 100-110A was needed for 4.0GHz. 110A*1.4V = 154W , which is much lower than what you would get by Anandtech's auto voltage per core numbers

Have you seen any other reviews comparing coolers on Ryzen?

► Recommended GPU Projects: [email protected] , [email protected] (FP64) (AMD moreso) ► Other notable GPU projects: [email protected] (Nvidia), GPUGrid (Nvidia) ► Project list


AlphaC is offline  
post #6 of 39 (permalink) Old 04-26-2017, 09:48 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 24,854
Rep: 1632 (Unique: 916)
Quote:
Originally Posted by AlphaC View Post

Well as far as the TDP goes, I've seen many sources (i.e. screenshots of hwinfo64 and buildzoid's testing) stating 100-110A was needed for 4.0GHz. 110A*1.4V = 154W , which is much lower than what you would get by Anandtech's auto voltage per core numbers

Have you seen any other reviews comparing coolers on Ryzen?
Yeah, I've seen a few showing the same 110A, but none for Ryzen yet.

We both know stock TDP has little to do with overclocked / high voltage heat.

Sadly I have not seen any Ryzen testing under coolers, but some cooler companies are just now getting a stock of AM4 mounts .. Sure would have been nice if AMD had released the offical mount specifications at least a few months ahead of Ryzen release instead of same day as release. mad.gif

I moved from AMD to Intel years ago, but Ryzen looks good .. might be time to move back. wink.gif

Phanteks Owners Club Ways to Better Cooling
i7 980 @ 3.55GHz =PH-TC14PE w/2x TY-143 fans =Crucial Ballistix 3x4GB =GA-X58A-UD5 =ASUS GTX580 DirectCU II =Enermax Modu84+ =Define R2 w/3x TY-140 case intake fans; all PWM controlled by CPU fan socketPhanteks Enthoo Primo MoBo Rampage III Extreme CPUi7 980X
@ 4.0GHz =R1 Ultimate w/2x TY-143 fans =Dominator GT 6x2GB =Rampage III Extreme =ASUS GTX580 DirectCU II =TX850 =Enthoo Primo w/ custom castor base
doyll is offline  
post #7 of 39 (permalink) Old 04-26-2017, 11:14 AM - Thread Starter
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 10,046
Rep: 815 (Unique: 548)
By the way , it's supposedly 110A for 4GHz all cores (i.e. overclocked to the so-called "safe" voltage wall). For XFR it's only on two cores.



and
Quote:
Originally Posted by https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/ 
On the high-end models the actual (effective) voltage for the base frequency (e.g. 3.6GHz on 1800X SKU) can be anything between 1.200 - 1.300V. Meanwhile the actual (effective) voltage for the highest single core boosted PState (XFR, e.g. 4.1GHz) can be as high as 1.47500V.
...
In the tested sample the actual default voltage for the base frequency (P0, 3.6GHz) was ~1.25000V, while the highest single core boost state (XFR, 4.1GHz) defaulted to 1.4625V.

► Recommended GPU Projects: [email protected] , [email protected] (FP64) (AMD moreso) ► Other notable GPU projects: [email protected] (Nvidia), GPUGrid (Nvidia) ► Project list


AlphaC is offline  
post #8 of 39 (permalink) Old 12-19-2017, 12:05 AM - Thread Starter
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 10,046
Rep: 815 (Unique: 548)
Interesting results on youtube with lower Ryzen CPUs

Quadcore

https://www.youtube.com/watch?v=FpjA3TdF6GE

Quad with SMT

https://www.youtube.com/watch?v=nrZnhdjLnGs

Hexcore with SMT



https://www.youtube.com/watch?v=1ZRnh3coMyw


https://www.youtube.com/watch?v=VNjkDoYZYjU

Cinebench R15 (~80W load) with Cryorig H7 Quad


https://www.youtube.com/watch?v=3OfFWetaWw0


https://www.youtube.com/watch?v=PP6vpl5Xuqs
Quote:
idle 41° - 44° (fan max speed) full load 69° - 73° (fan max speed) hardware: CPU RYZEN 5 1600 OC @3.6Ghz 1.3v Motherboard MSI B350M GAMING PRO Memory Patriot Viper 4 16GB 3400MHz (running 3200MHz)
Prime95 FMA3



Ryzen 7

https://techbuyersguru.com/cpu-liquid-cooler-shootout-pushing-limits?page=2



(load at wall)

Intro to the Arctic, Cryorig, Noctua, Scythe & SilverStone Ryzen Air Cooler Shootout ... all talk and no data yet. rolleyes.gif
https://www.youtube.com/watch?v=L44WvIBSQU0

► Recommended GPU Projects: [email protected] , [email protected] (FP64) (AMD moreso) ► Other notable GPU projects: [email protected] (Nvidia), GPUGrid (Nvidia) ► Project list


AlphaC is offline  
post #9 of 39 (permalink) Old 12-30-2017, 04:08 PM
Linux Lobbyist
 
2overclock's Avatar
 
Join Date: Jul 2017
Posts: 71
Rep: 2 (Unique: 2)
I have a Ryzen 5 1600 OC 3.9 @ 1.5V, corsair ram 16g @ 3200mhz on a msi b350 gaming pro mobo. I used the stock heatsink for a while then switched to deepcool gammaxx 400 cpu heatsink with notcua's 3000 rpm 120mm fan and I played bf1 for hours on a 1080 144mhz @ 1440p, and the highest temp on the package has reached 71C. I just wanted to share my experience.
2overclock is offline  
post #10 of 39 (permalink) Old 12-31-2017, 05:46 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 24,854
Rep: 1632 (Unique: 916)
Quote:
Originally Posted by 2overclock View Post

I have a Ryzen 5 1600 OC 3.9 @ 1.5V, corsair ram 16g @ 3200mhz on a msi b350 gaming pro mobo. I used the stock heatsink for a while then switched to deepcool gammaxx 400 cpu heatsink with notcua's 3000 rpm 120mm fan and I played bf1 for hours on a 1080 144mhz @ 1440p, and the highest temp on the package has reached 71C. I just wanted to share my experience.
Was cooler fan at full speed?
What case and case fans do you have and their speeds?
How loud is our system?

I ask because I've tested many coolers with stock and high speed fans like NF-A12 iPPC 3000rpm an found the difference on a high heat CPU was 8-10c, which is niice .. but the noise level to get the 8-10c better cooling was 4-8 times louder, which to me was way too loud, so I run fans at lower speed and lower the overclock 200-400MHz and ran 3.8-4.0GHz instead of 4.2GHz. I didn't notice any difference in performance, even on long encoding sessions it was only a few minutes in an hour or so of encoding. But then I don't like a lot of noise. biggrin.gif

Phanteks Owners Club Ways to Better Cooling
i7 980 @ 3.55GHz =PH-TC14PE w/2x TY-143 fans =Crucial Ballistix 3x4GB =GA-X58A-UD5 =ASUS GTX580 DirectCU II =Enermax Modu84+ =Define R2 w/3x TY-140 case intake fans; all PWM controlled by CPU fan socketPhanteks Enthoo Primo MoBo Rampage III Extreme CPUi7 980X
@ 4.0GHz =R1 Ultimate w/2x TY-143 fans =Dominator GT 6x2GB =Rampage III Extreme =ASUS GTX580 DirectCU II =TX850 =Enthoo Primo w/ custom castor base
doyll 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)
 


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