Zen2 waterblock comparison - Overclock.net - An Overclocking Community

Forum Jump: 

Zen2 waterblock comparison

Reply
 
Thread Tools
post #1 of 38 (permalink) Old 07-28-2019, 01:49 AM - Thread Starter
Wubbalubba Dub Dub
 
CiBi's Avatar
 
Join Date: Jan 2008
Location: Belgium
Posts: 961
Rep: 21 (Unique: 19)
Zen2 waterblock comparison

Hey guys

It might be informative for everyone to post their thermal results at stock settings for Zen2 along with the cooling configuration used.
I get the impression that some blocks cool the off centre die better than others.

Nikon D610 // Nikkor 24-120mm F4 // Nikkor 18-35mm F3.5-4.5 // Tokina 105mm F2.8 // Voightlander 40mm F2.0
Ants In My Skylake
(14 items)
Ants In My HTPC
(8 items)
CPU
Intel i7-6700K (4.60Ghz @ 1.35v)
Motherboard
Asus Z170 Pro
GPU
Asus GTX 1070 8GB
RAM
32GB Kingston HyperX Fury DDR4 2133Mhz
Hard Drive
Samsung 850 SSD (250GB)
Hard Drive
WD 3TB
Power Supply
XFX PRO 550W Core Edition
Cooling
Corsair H110i
Case
NZXT Switch 810 (white)
Operating System
Windows 10 64bit
Monitor
AOC 27" QHD 144Hz
Keyboard
Corsair K70 (Cherry MX Red)
Mouse
Razer DeathAdder Elite
Audio
Bose QC25
CPU
AMD Phenom II X6 1090T (3.50Ghz @ 1.37v)
Motherboard
AMD Sabertooth 990FX
GPU
MSI GTX 770 2GB
RAM
8GB Corsair Vengeance DDR3 1333Mhz
Hard Drive
Samsung 830 (128GB)
Power Supply
Corsair VS550
Cooling
Cooler Master Hyper 212X
Operating System
Windows 10 64bit
▲ hide details ▲


CiBi is offline  
Sponsored Links
Advertisement
 
post #2 of 38 (permalink) Old 07-28-2019, 02:06 AM
New to Overclock.net
 
andrewmp6's Avatar
 
Join Date: Oct 2017
Location: Louisville Kentucky
Posts: 541
Rep: 15 (Unique: 14)
Heatkiller will cool and flow the best of any cpu block known.
andrewmp6 is offline  
post #3 of 38 (permalink) Old 07-28-2019, 11:15 AM
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 11,185
Rep: 908 (Unique: 590)
Zen 2 acts like a GPU , it boosts based on temperature and power. Unless you control the power limit then the results would be completely all over the place.

For example a Ryzen 5 3600 on stock Wraith Stealth cooler might reach only base 3.6GHz at 91°C in Prime95 Small FFTs on stock settings but something like an Eisblock XPX (overkill for R5 3600) would reach 65°C with ~3.8GHz.

Also you would have a discrepancy between Ryzen 5 or 7 which is one die versus Ryzen 9 which has two of them.

► 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 38 (permalink) Old 07-29-2019, 03:59 AM - Thread Starter
Wubbalubba Dub Dub
 
CiBi's Avatar
 
Join Date: Jan 2008
Location: Belgium
Posts: 961
Rep: 21 (Unique: 19)
Quote: Originally Posted by AlphaC View Post
Zen 2 acts like a GPU , it boosts based on temperature and power. Unless you control the power limit then the results would be completely all over the place.

For example a Ryzen 5 3600 on stock Wraith Stealth cooler might reach only base 3.6GHz at 91°C in Prime95 Small FFTs on stock settings but something like an Eisblock XPX (overkill for R5 3600) would reach 65°C with ~3.8GHz.

Also you would have a discrepancy between Ryzen 5 or 7 which is one die versus Ryzen 9 which has two of them.
Thanks for the reply, you are obviously correct, at stock we can't really compare.
Would be nice to see a test of several blocks at fixed voltages en frequencies because when looking around I notice some blocks/aio's cool way better then others. Probably because of how wide the microfins extend inside the block.

Nikon D610 // Nikkor 24-120mm F4 // Nikkor 18-35mm F3.5-4.5 // Tokina 105mm F2.8 // Voightlander 40mm F2.0
Ants In My Skylake
(14 items)
Ants In My HTPC
(8 items)
CPU
Intel i7-6700K (4.60Ghz @ 1.35v)
Motherboard
Asus Z170 Pro
GPU
Asus GTX 1070 8GB
RAM
32GB Kingston HyperX Fury DDR4 2133Mhz
Hard Drive
Samsung 850 SSD (250GB)
Hard Drive
WD 3TB
Power Supply
XFX PRO 550W Core Edition
Cooling
Corsair H110i
Case
NZXT Switch 810 (white)
Operating System
Windows 10 64bit
Monitor
AOC 27" QHD 144Hz
Keyboard
Corsair K70 (Cherry MX Red)
Mouse
Razer DeathAdder Elite
Audio
Bose QC25
CPU
AMD Phenom II X6 1090T (3.50Ghz @ 1.37v)
Motherboard
AMD Sabertooth 990FX
GPU
MSI GTX 770 2GB
RAM
8GB Corsair Vengeance DDR3 1333Mhz
Hard Drive
Samsung 830 (128GB)
Power Supply
Corsair VS550
Cooling
Cooler Master Hyper 212X
Operating System
Windows 10 64bit
▲ hide details ▲


CiBi is offline  
post #5 of 38 (permalink) Old 07-29-2019, 10:59 AM
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 11,185
Rep: 908 (Unique: 590)
Fixed voltage and frequencies would have to account for differences in CPU quality (leakage). The only proper way of comparing would be to use power density, which means the power has to be a fixed amount per die. At least then you have a fixed variable.

Ultimately it's only worth comparing on Ryzen 9 chips because realistically there isn't a good reason to watercool a $200 Ryzen 5 , and watercooling the $330-400 Ryzen 7 is debatable. A decent full copper kit is $150-300 depending on pump quality , so unless by watercooling you mean $70 240mm bargain bin AIOs there isn't a price/performance reason to watercool the low TDP chips. If we go by past Ryzen chips, you're looking at ~20-22W per core and the only time that would make sense to watercool is when you have 10+ cores. 8 cores would have you at ~180W which is just above the TDP rating of a decent 4 heatpipe cooler and a 6 heatpipe cooler would give you enough thermal headroom with reasonable fan speeds.


Peak power loading for Ryzen 3rd gen seems to be around 16-18W per core , which makes this even more true. (https://www.anandtech.com/show/14605...ing-the-bar/19)


Power:
Quote: Originally Posted by https://www.overclockers.com/amd-ryzen-9-3900x-and-ryzen-7-3700x-cpu-review/
In the graph below we tested power use of the system across multiple situations from idle, to Prime 95 Small FFT (with FMA3/AVX). The 3900X pulled the most power during the AIDA64 FPU test with 220 W. What’s interesting, that’s 5 W lower than the 2700X in our previous testing despite adding 50% more cores. This really shows the benefits of the 7 nm process and AMD’s efficiency. The 3700X seemed to pull the same power from the wall in all tests at a steady 160 W (we are not sure why at this point and will update this section as information comes in). Keep in mind this is full-system power usage.

...
R7 3700X @ 4.35GHz = 76.8°C peak temperature / 59°C average , 107.37W peak package power

I switched over to my EKWB Predator 360 XLC closed-loop cooler for improved cooling which took me up to 4.35 GHz and stable on both the Ryzen 7 3700X with 1.45 V and the Ryzen 9 3900X with 1.4 V.

R9 3900X @4.35GHz = 104.1°C peak temperature , 246W peak package power

Blender with Noctua NH-U12 (5 heatpipe cooler with 120mm fan):
https://www.techpowerup.com/review/a...-3900x/19.html
R7 3700X = 66°C
R9 3900X = 79°C

Japanese testing, closed loop Corsair vs wraithripper (an odd choice of cooler for Ryzen):
Click image for larger version

Name:	41_980x640.jpg
Views:	27
Size:	64.2 KB
ID:	284376
Click image for larger version

Name:	40_980x641.jpg
Views:	30
Size:	52.6 KB
ID:	284378
https://ascii.jp/elem/000/001/891/1891699/index-8.html


See also https://www.techspot.com/review/1875...liquid-cooler/

► 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 38 (permalink) Old 07-29-2019, 11:22 AM
New to Overclock.net
 
jtom320's Avatar
 
Join Date: Oct 2011
Location: Washington DC
Posts: 3,794
Rep: 128 (Unique: 109)
I have a Heatkiller IV on a 3900x.

20 - 22 degrees C room temps.

I see with PBO and "advanced" overlocking enabled with the 200mhz offset temps still in the 70 degree range on things like cinebench or general gaming (Mostly CSGO so grain of salt there). I've let small FFTs run for about 8 hours just to see where it'd max out at and I believe 77 degrees was the max recorded core temp when I got home for the day.

But I also think this chip is a bios update away from behaving properly because even when I'm running like a single window of chrome the thing is gulping voltage. At least according to CPUID which was recommended by the AMD guy over on Reddit.

I've tried a bit of manual overclocking and would be happy to provide some data. Rad size shouldn't be a problem I'm running two 360's and an idle 1080.


Paste is Thermal Grizzly.

Nvidia Sparse Grid Supersampling guide.
Ryzen 3000
(17 items)
CPU
R9 3900x
Motherboard
Gigabyte Master
GPU
GTX 1080
RAM
32gb DDR4 3433 CL16
Hard Drive
Samsung 950 Pro 256gb
Hard Drive
Samsung 850 Evo 512gb
Hard Drive
Samsung 850 Evo 512gb
Power Supply
Seasonic Focus Plus 850 Gold
Cooling
HeatKiller IV AM4
Cooling
EK 1080 FE GPU Block
Cooling
Black Ice 360mm SR2
Cooling
Black Ice Nemesis 360mm
Cooling
Aquacomputer Aqualis Fountain Thing
Cooling
D5 Vario
Cooling
PETG
Case
Fractal S2
Operating System
Win 10
CPU
6700k (EK Supremacy Evo) *4.7 @ 1.312
Motherboard
Asus Maximus VIII gene
GPU
GTX 1080 FE (EK nickel/plexi)* 2075/5401
RAM
16gb Corsair 3200
Hard Drive
Samsung 950pro M.2 256gb
Hard Drive
Samsung 850 Evo - 512 - Steam
Hard Drive
Samsung 850 Evo - 512 - Other Apps
Power Supply
EVGA 650
Cooling
EK Res
Cooling
EK pump top
Cooling
XSPC EX 360
Cooling
XSPC EX 240
Case
Fractal Arc Mini 2
Operating System
10
Keyboard
Ducky Shine 3 TKL
Mouse
G303
▲ hide details ▲
jtom320 is offline  
post #7 of 38 (permalink) Old 07-29-2019, 03:16 PM
New to Overclock.net
 
japanesegorilla's Avatar
 
Join Date: Feb 2012
Location: Gresham, OR
Posts: 224
Rep: 16 (Unique: 16)
At 22C ambient, under Prime95 smallest FFTs, stock cooler my 3600 was hitting 95C 3.9GHz at 1.45V but would eventually throttle down to 3.6GHz all cores as it lowered voltage to try to reduce temperatures.

At 22C ambient, under Prime95 smallest FFTs, XSPC Raystorm (original, not pro or neo), my 3600 hits 76C 4.1GHz at 1.35V and roughly maintains that frequency. I should note that this was after I greatly tightened down the corner screw nearest to the corner chiplet on the 3600. I don't know if the Raystorm was just not making good contact with the IHS of the 3600 or if putting more pressure on that corner helps because there is a chiplet under that corner. Before tightening that corner all the way down (while leaving the others a few turns looser) I was still hitting 95C under full water.

Intel I5-2500K (@ 4.5GHz), Asrock P67 Extreme4 Gen3, 16GB G.Skill DDR3 1600 F3-12800CL9D-8GBXL @ 8-8-8-20, EVGA GTX 980Ti SC+, Thermaltake Core X71, XSPC Raystorm CPU, EX 120, EX360, EX480, MCP35X
japanesegorilla is offline  
post #8 of 38 (permalink) Old 08-01-2019, 09:53 AM - Thread Starter
Wubbalubba Dub Dub
 
CiBi's Avatar
 
Join Date: Jan 2008
Location: Belgium
Posts: 961
Rep: 21 (Unique: 19)
Quote: Originally Posted by AlphaC View Post
Fixed voltage and frequencies would have to account for differences in CPU quality (leakage). The only proper way of comparing would be to use power density, which means the power has to be a fixed amount per die. At least then you have a fixed variable.

Ultimately it's only worth comparing on Ryzen 9 chips because realistically there isn't a good reason to watercool a $200 Ryzen 5 , and watercooling the $330-400 Ryzen 7 is debatable. A decent full copper kit is $150-300 depending on pump quality , so unless by watercooling you mean $70 240mm bargain bin AIOs there isn't a price/performance reason to watercool the low TDP chips. If we go by past Ryzen chips, you're looking at ~20-22W per core and the only time that would make sense to watercool is when you have 10+ cores. 8 cores would have you at ~180W which is just above the TDP rating of a decent 4 heatpipe cooler and a 6 heatpipe cooler would give you enough thermal headroom with reasonable fan speeds.


Peak power loading for Ryzen 3rd gen seems to be around 16-18W per core , which makes this even more true. (https://www.anandtech.com/show/14605...ing-the-bar/19)


Power:



Blender with Noctua NH-U12 (5 heatpipe cooler with 120mm fan):
https://www.techpowerup.com/review/a...-3900x/19.html
R7 3700X = 66°C
R9 3900X = 79°C

Japanese testing, closed loop Corsair vs wraithripper (an odd choice of cooler for Ryzen):
Attachment 284376
Attachment 284378
https://ascii.jp/elem/000/001/891/1891699/index-8.html


See also https://www.techspot.com/review/1875...liquid-cooler/
First off, thanks for the very elaborate post, a lot of information there.
Lets all agree watercooling is rarely a necessity, even with the 3900X

I'm looking to cool a 3800X at however fast i can make it run
I eventually bit the bullet and loaded up my virtual cart with 530€ worth of watercooling gear. That is excluding the block for the GPU but including fittings radspace and fans for the GPU. Mostly EK stuff, noctua fans, D5 pump with glass res.

I'll see to it that I post some more data regarding my thermals here, however relevant it might be with the boosting and all.
I'll probably also get a build log started again soon.

Quote: Originally Posted by japanesegorilla View Post
At 22C ambient, under Prime95 smallest FFTs, stock cooler my 3600 was hitting 95C 3.9GHz at 1.45V but would eventually throttle down to 3.6GHz all cores as it lowered voltage to try to reduce temperatures.

At 22C ambient, under Prime95 smallest FFTs, XSPC Raystorm (original, not pro or neo), my 3600 hits 76C 4.1GHz at 1.35V and roughly maintains that frequency. I should note that this was after I greatly tightened down the corner screw nearest to the corner chiplet on the 3600. I don't know if the Raystorm was just not making good contact with the IHS of the 3600 or if putting more pressure on that corner helps because there is a chiplet under that corner. Before tightening that corner all the way down (while leaving the others a few turns looser) I was still hitting 95C under full water.
I'll have to experiment with the screw pressure as well I guess, 95C under full water is a bad time. I superimposed naked zen2 die pics on top of waterblocks to see which one had some microfin coverage over the actual die. Most of them do, its very off centre though.
I asked EK and they informed me they were testing if alternative waterblock design was needed. So could be a while before they release a block if they even release a new one at all.
I eventually settled on an EK Velocity, i'm a bit biased towards EK...

Nikon D610 // Nikkor 24-120mm F4 // Nikkor 18-35mm F3.5-4.5 // Tokina 105mm F2.8 // Voightlander 40mm F2.0
Ants In My Skylake
(14 items)
Ants In My HTPC
(8 items)
CPU
Intel i7-6700K (4.60Ghz @ 1.35v)
Motherboard
Asus Z170 Pro
GPU
Asus GTX 1070 8GB
RAM
32GB Kingston HyperX Fury DDR4 2133Mhz
Hard Drive
Samsung 850 SSD (250GB)
Hard Drive
WD 3TB
Power Supply
XFX PRO 550W Core Edition
Cooling
Corsair H110i
Case
NZXT Switch 810 (white)
Operating System
Windows 10 64bit
Monitor
AOC 27" QHD 144Hz
Keyboard
Corsair K70 (Cherry MX Red)
Mouse
Razer DeathAdder Elite
Audio
Bose QC25
CPU
AMD Phenom II X6 1090T (3.50Ghz @ 1.37v)
Motherboard
AMD Sabertooth 990FX
GPU
MSI GTX 770 2GB
RAM
8GB Corsair Vengeance DDR3 1333Mhz
Hard Drive
Samsung 830 (128GB)
Power Supply
Corsair VS550
Cooling
Cooler Master Hyper 212X
Operating System
Windows 10 64bit
▲ hide details ▲


CiBi is offline  
post #9 of 38 (permalink) Old 08-01-2019, 08:29 PM
New to Overclock.net
 
PhillyB's Avatar
 
Join Date: Sep 2014
Location: Kincheloe, MI
Posts: 491
Rep: 26 (Unique: 25)
Running a 3900x on a x470 Pro Carbon with an EK monoblock. 2x XSPC RX360 with EK Vardar 1150rpm.

At last test room was about 22-23c, CPU peaked at 71c in a cinebench20 run. CPU all core 4.3Ghz at 1.285v.

The current bios versions pour voltage into the 3000 series at ridiculous levels at stock. The benefit of overclocking is limited with these chips for performance, but just doing it to maintain reasonable voltages makes it almost a necessity in my opinion. There was minor reduction in temps at each step downward from 1.32v. On average, temps are staying in the 50s on normal workloads.

I tried dropping volts without removing stock core power table settings. It just down clocked the CPU well below base clock when at 1.32v. Considering the chip maintains 4.3ghz at 1.285v the default bios settings are complete crap right now.

Did I mention that I also like consistent performance out of my computer...
PhillyB is offline  
post #10 of 38 (permalink) Old 08-02-2019, 01:23 AM
New to Overclock.net
 
mongoled's Avatar
 
Join Date: Oct 2011
Location: Ayia Napa, Cyprus
Posts: 481
Rep: 17 (Unique: 13)
Quote: Originally Posted by japanesegorilla View Post
At 22C ambient, under Prime95 smallest FFTs, stock cooler my 3600 was hitting 95C 3.9GHz at 1.45V but would eventually throttle down to 3.6GHz all cores as it lowered voltage to try to reduce temperatures.

At 22C ambient, under Prime95 smallest FFTs, XSPC Raystorm (original, not pro or neo), my 3600 hits 76C 4.1GHz at 1.35V and roughly maintains that frequency. I should note that this was after I greatly tightened down the corner screw nearest to the corner chiplet on the 3600. I don't know if the Raystorm was just not making good contact with the IHS of the 3600 or if putting more pressure on that corner helps because there is a chiplet under that corner. Before tightening that corner all the way down (while leaving the others a few turns looser) I was still hitting 95C under full water.
Hi!

What version of prime did you run ??

Look at my sig (custom loop), when running prime95 29.8 build 5, PBO on auto, 3800/1900 mhz, using blend test I see peaks of almost 85C.

Does the version you use have AVX2 ??

Quote: Originally Posted by AlphaC View Post
Ultimately it's only worth comparing on Ryzen 9 chips because realistically there isn't a good reason to watercool a $200 Ryzen 5 , and watercooling the $330-400 Ryzen 7 is debatable. A decent full copper kit is $150-300 depending on pump quality , so unless by watercooling you mean $70 240mm bargain bin AIOs there isn't a price/performance reason to watercool the low TDP chips. If we go by past Ryzen chips, you're looking at ~20-22W per core and the only time that would make sense to watercool is when you have 10+ cores. 8 cores would have you at ~180W which is just above the TDP rating of a decent 4 heatpipe cooler and a 6 heatpipe cooler would give you enough thermal headroom with reasonable fan speeds.
Although im in agreement, the more cooler these CPU run, the more consistent and better performing the results are.

Also those of us who already have custom loops we will of course still use them with any CPU



Now

Ive attached a screen shot from a 12+ hr prime95 blend. You can see the relavant information

Im also attaching a picture (the 'ridgy' area is a overlay of how the conductonaut spread) of how my waterblocks head sits on the IHS and its position relative to the CPU and I/O die.

I contacted EK regards the fins positioning but from their answer seems like they are not interested in redesigning a different head for my monoblock



Im very interested to see more results in this thread, hope others will participate soon

Attached Thumbnails
Click image for larger version

Name:	prime95-blend_3800-1900-tCL16-tRCDRD15- tRCDWR15-tRP15-tRAS30- tRC52-tRFC303-tFAW36- tWR24.png
Views:	59
Size:	413.3 KB
ID:	285210  

Attached Images
 

SeaSonic X750 Gold 750W
MSI X370 XPOWER Gaming Titanium | Ryzen 5 3600X 1924SUT
GSkill TridentZ F4-3600C15D-16GTZ @ 3800/1900 mhz | 1.44v | 15-17-10-15-42-28-266-16-1T
Asus GTX 780 DCII OC @ 1332/3522 mhz | 1.3v | Samsung 960 EVO 500GB
EK-FB MSI X370 XPower RGB Monoblock 1/2" ID | Swiftech D5 MCP655 | EK-XRES 100
Alphacool NexXxoS UT60 280 | 4 x Scythe GT AP-15 @ 5v
Thermochill PA120.3 | 1 x SST-FN123 | 2 x Scythe GT AP-15 | @ 5v

Last edited by mongoled; 08-02-2019 at 01:36 AM.
mongoled 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