Threadripper BUILDERS thread

Page 28 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

hasta666

Junior Member
Aug 18, 2017
17
8
81
Depending on when my MB arrives, I'll consider the Noctua heatsink. I'm not really optimistic about the availability of the LiqTech TR, especially in my country.

In the mean time, as I'm waiting on literally everything I ordered, I bought a new case and a new PSU in order to avoid the hassle of unmounting the old components/mounting the new ones.
 

StefanR5R

Elite Member
Dec 10, 2016
5,690
8,263
136
OK, How is this revised list ?
Looks good to me. If you want to expand to GPU watercooling later, (1.) you could consider a thicker radiator now of course, like others already commented, and (2.) when you order additional parts for GPU watercooling, then order plenty of tubing length, i.e. more than you need immediately. This will allow you to try and redo loop layouts without having to order more tubes and wait for them to arrive.

My own watercooled PC (which is not for looks, only for quietness while having lots of GPU power concentrated in a single host) is with 10/16 EPDM tubing. The thicker walls of 10/16 tubes are said to have a lesser tendency to kink in bends, and EPDM is said to be even more resistant to deterioration than PVC tubing. But I can't compare myself as I haven't handled other tubing myself yet.

The XE in push-pull @ 1300rpm will exhaust around 360W on a 10C gradient, while the Nemesis 360 GTS (a slim radiator like the SE) will exhaust only 309W. So yeah, if it fits....
http://www.xtremerigs.net/2015/05/31/ek-coolstream-xe-360mm-radiator-review/5/
EK also published some comparisons between their own radiator models:
https://www.ekwb.com/blog/radiators-part-1-thickness-fpi/
https://www.ekwb.com/blog/radiators-part-2-performance/
https://www.ekwb.com/blog/radiators-part-3-surface-thickness/

The only thing I saw missing on your list (that I use) was a mini-valve for draining your loop. An example is here http://www.performance-pcs.com/bitspower-bp-mvv-mbk-mini-valve-matt-black.html . You'll need at least 1 stop plug for that.
I simply use the reservoir to drain the loop. This method is facilitated by the absence of any aesthetic requirements on my PC...
 
Reactions: dnavas and IEC

IEC

Elite Member
Super Moderator
Jun 10, 2004
14,362
5,027
136
My own watercooled PC (which is not for looks, only for quietness while having lots of GPU power concentrated in a single host) is with 10/16 EPDM tubing. The thicker walls of 10/16 tubes are said to have a lesser tendency to kink in bends, and EPDM is said to be even more resistant to deterioration than PVC tubing. But I can't compare myself as I haven't handled other tubing myself yet.

I use 10mm/16mm (3/8" ID, 5/8" OD for us Imperials) Tygon A-60-G Norprene tubing for the same reasons. It doesn't kink easily, it doesn't leach plasticizer, and it is actually used in industrial and laboratory applications with a good track record.
 

ub4ty

Senior member
Jun 21, 2017
749
898
96
FWIW, the Enermax liqtech was on sale briefly at newegg earlier in the day (it's again out of stock). I don't fancy mixing copper and aluminum myself, but presumably someone will be in a position to judge its cooling effectiveness soon.
What do you mean by copper mixing with aluminum? all AIOs have copper contact blocks... I don't seem to understand what has ever been the issue w/ that...
 
Reactions: Drazick

ddogg

Golden Member
May 4, 2005
1,864
361
136
I simply use the reservoir to drain the loop. This method is facilitated by the absence of any aesthetic requirements on my PC...

I don't suggest this method especially if you have a complex loop and a big heavy case. Installing a drain port makes life very easy to drain the loop.
 

dnavas

Senior member
Feb 25, 2017
355
190
116
What do you mean by copper mixing with aluminum? all AIOs have copper contact blocks... I don't seem to understand what has ever been the issue w/ that...

First, I don't know whether the copper is actually in contact with the coolant in the liqtech, and plenty of people will tell you that with the right anti-corrosive additives, you'll be "fine". But, I keep my computers for 7-year stretches, and I'm skeptical over that timeframe if the copper is in contact with the coolant, and if it's not, there are likely thermal expansion issues.

For electric wiring, you should be told to not connect aluminum wiring to copper contacts (and vice versa) because the thermal expansion of the two metals is different, and the wiring will eventually work itself loose, spark, and cause fires. However, for cooling loops, the issue is different and even easier to understand: two different metals + ion carrier (like water) == battery (galvanic cell, hence galvanic corrosion). One metal will essentially dissolve, and really, I don't want my aluminum radiator dissolving, that would make the motherboard under it unhappy.

Here's more reading if you'd like: https://www.ekwb.com/blog/aluminium-vs-copper/

random trivia: aluminum was once more precious than gold, aluminum oxide can be quite pretty (think rubies and sapphires) and spaceX lost their first Falcon 1 due to a corroded aluminum nut (it was replaced with a steel one -- steel + aluminum == unhappiness, and the rocket was exposed to sea spray in Kwajalein).
 
Reactions: IEC and Yakk

Chaotic42

Lifer
Jun 15, 2001
33,929
1,098
126
This is the thread for that !!

So it looks like I'm the victim of two sets of failures:

1) My orange carrier frame is broken. There are six little clips at the bottom. Only three have the edges to hold the CPU in. The others are just cut off. It felt really janky when I opened it, I guess that's why

2) My motherboard is suffering from screws which are too short. None of the star screws on the CPU latch bite. This is a known problem: https://community.amd.com/thread/219286

I'd like to have a positive experience with AMD at least once in my life.
 

IEC

Elite Member
Super Moderator
Jun 10, 2004
14,362
5,027
136
So it looks like I'm the victim of two sets of failures:

1) My orange carrier frame is broken. There are six little clips at the bottom. Only three have the edges to hold the CPU in. The others are just cut off. It felt really janky when I opened it, I guess that's why

2) My motherboard is suffering from screws which are too short. None of the star screws on the CPU latch bite. This is a known problem: https://community.amd.com/thread/219286

I'd like to have a positive experience with AMD at least once in my life.

No idea about #1, but #2 is an issue with some Foxconn TR4 socket retention mechanisms (maybe exclusive to ASUS boards?). Is your socket made by Foxconn? My MSI board has one made by LOTES and I didn't have any issues.
 
Reactions: Drazick

ub4ty

Senior member
Jun 21, 2017
749
898
96
First, I don't know whether the copper is actually in contact with the coolant in the liqtech, and plenty of people will tell you that with the right anti-corrosive additives, you'll be "fine". But, I keep my computers for 7-year stretches, and I'm skeptical over that timeframe if the copper is in contact with the coolant, and if it's not, there are likely thermal expansion issues.

For electric wiring, you should be told to not connect aluminum wiring to copper contacts (and vice versa) because the thermal expansion of the two metals is different, and the wiring will eventually work itself loose, spark, and cause fires. However, for cooling loops, the issue is different and even easier to understand: two different metals + ion carrier (like water) == battery (galvanic cell, hence galvanic corrosion). One metal will essentially dissolve, and really, I don't want my aluminum radiator dissolving, that would make the motherboard under it unhappy.

Here's more reading if you'd like: https://www.ekwb.com/blog/aluminium-vs-copper/

random trivia: aluminum was once more precious than gold, aluminum oxide can be quite pretty (think rubies and sapphires) and spaceX lost their first Falcon 1 due to a corroded aluminum nut (it was replaced with a steel one -- steel + aluminum == unhappiness, and the rocket was exposed to sea spray in Kwajalein).

But were talking about AIO coolers not custom loops. You mentioned enermax's cooler and said you were worried about copper and aluminum galvanizing.. Every AIO would face this issue and I'm sure the sea of AIO cooler manufacturers thought about that and engineered their systems to protect against that. The cooler on Vega isn't going to suddenly begin galvanized/etc. All AIOs have this design. If I am incorrect and they all galvanize over time and are useless, its yet another reason for me not to even waste my time/$$$/risk my setup for something that marginally performs better/worse than air.

Anywho, I just placed an order for the Noctua cooler found here :
https://www.newegg.com/Product/Product.aspx?Item=9SIAADY62M6046

Even though reviewers have had it for approx a week, it seems they're dragging tail. There was one good review at stock and the numbers looked good enough for me.

I never had any desire to bring water near such a workstation and all of the research and chatter needed w.r.t to water cooling made my head spin. Per one review, it managed to beat the thermaltake floring 360 in cooling performance and has a full contact plate. No risk of dealing w/ water issues and the Noctua is only $80. I'd want an air cooler as backup anyway. So, I have it now and will likely only look at water if there's a need. It appears the noctua holds good temps at stock w/ boost to 4.2Ghz. A user mentioned they regular see it do 3.7/3.8ghz boost on all cores. I really have no desire to dump $$$ to get 200-400mhz more. AIO don't seem to be the route and were talking $250-$400 for a custom loop. So...


$80 and air it is. ez pz
 

dnavas

Senior member
Feb 25, 2017
355
190
116
But were talking about AIO coolers not custom loops. [...] All AIOs have this design. If I am incorrect and they all galvanize over time and are useless, its yet another reason for me not to even waste my time/$$$/risk my setup

Yeah, my suspicion is that the coolant will evaporate/leak faster (the tubing is generally not great on closed-loop AIOs as well -- they're cheap for a reason :sigh: ).

I never had any desire to bring water near such a workstation

Yeah, I'm not thrilled either, but I can definitely use the extra Mhz, and I'm hoping to benefit from noise reduction with lower fan speeds. Been an interesting learning experience, anyway. Maybe Noctua will release an air cooler with a little more OC space in a couple of months.
 

ub4ty

Senior member
Jun 21, 2017
749
898
96
Yeah, my suspicion is that the coolant will evaporate/leak faster (the tubing is generally not great on closed-loop AIOs as well -- they're cheap for a reason :sigh: ).



Yeah, I'm not thrilled either, but I can definitely use the extra Mhz, and I'm hoping to benefit from noise reduction with lower fan speeds. Been an interesting learning experience, anyway. Maybe Noctua will release an air cooler with a little more OC space in a couple of months.
You're likely right.. that seems to be the end-state of AIOs after some time which has always concerned me as they're black boxes purposely made to prevent maintenance or deconstruction. Meanwhile, prices skyrocket because water cooling is considered to be a luxury market.

The Noctua beat a 360 rad on temps that people are OCing on. I'm sure it will handle 3.7/3.8 24/7 operation and 4.0Ghz infrequent operation. I'm not seeing the leap in cost justification for 100mhz or 200mhz. I also don't see someone saying their life depends on 100/200mhz speed diff. I do see someone saying they want to pursue for hobby/aesthetics but then again that's why Water cooling is so pricy.. It's a luxury hobby segment of the market. That being said, let me shut up about it. I don't want to ruin anyone's hobby . I look at such setups and I get mildly hard. Thus, i can't deny the value in that. I will hide my poo colored Noctua and drool over your water loops.
 

Chaotic42

Lifer
Jun 15, 2001
33,929
1,098
126
Holy cow. So much force required to get that plate screwed in, my carrier frame just totally failed, my motherboard's little display on the inside is all broken... but it does POST. Complete mess, but it's there. Now to wait for the joyous AT&T internet to download Ubuntu 17.04...
 

ub4ty

Senior member
Jun 21, 2017
749
898
96
Holy cow. So much force required to get that plate screwed in, my carrier frame just totally failed, my motherboard's little display on the inside is all broken... but it does POST. Complete mess, but it's there. Now to wait for the joyous AT&T internet to download Ubuntu 17.04...
Ubuntu 17.04.
What plans do you have for this build my friend?
Something smells good ...
 
Reactions: Drazick

eek2121

Diamond Member
Aug 2, 2005
3,051
4,276
136
I definitely need a better cooler for this beast. I can OC to 4GHz just fine, but applying the types of load that P95 or 'stress CPU' of CPU-Z applies sends temps through the roof. According to Ryzen Master the temps got over 80C with
You're likely right.. that seems to be the end-state of AIOs after some time which has always concerned me as they're black boxes purposely made to prevent maintenance or deconstruction. Meanwhile, prices skyrocket because water cooling is considered to be a luxury market.

The Noctua beat a 360 rad on temps that people are OCing on. I'm sure it will handle 3.7/3.8 24/7 operation and 4.0Ghz infrequent operation. I'm not seeing the leap in cost justification for 100mhz or 200mhz. I also don't see someone saying their life depends on 100/200mhz speed diff. I do see someone saying they want to pursue for hobby/aesthetics but then again that's why Water cooling is so pricy.. It's a luxury hobby segment of the market. That being said, let me shut up about it. I don't want to ruin anyone's hobby . I look at such setups and I get mildly hard. Thus, i can't deny the value in that. I will hide my poo colored Noctua and drool over your water loops.

I'm going to have to interject here. Most AIOs aren't designed for Threadripper. They don't have direct contact with Threadripper, the contact surface of the cooler is concave, not flat. I originally read this on hardocp, but didn't think it was that big of a deal until I removed my cooler and found that there was very little thermal paste in the middle. I applied additional paste in the middle and put everything back together. Now I'm getting much cooler temps...62 degrees @ 3.8 (100% load) and 1.275 volts on a 240 mm rad pushing out warm air from the case (due to the lack of needed hardware this setup was necessary). I will try 4 ghz tomorrow some time, but so far things are looking great. I know voltages can go lower, so I'm betting on a good OC.
 

Chaotic42

Lifer
Jun 15, 2001
33,929
1,098
126
Okay. Sweet Fancy Moses.

I'll update this post tomorrow when I have a lot less hate for anything computer related. If someone is trying to get the wifi going on their ROG Zenith Extreme in linux, here's what worked for me. The firmware for the device in the distros seems to be broken. I followed some of the steps from this site and have listed them below (https://wiki.gentoo.org/wiki/Qualcomm_Atheros_QCA6174)

-Edit your /etc/NetworkManager/NetworkManager.conf file
-Change managed=false to managed=true
-Shut down your computer and start it back up
-Then run the following

Code:
sudo apt-get install linux-firmware
mv /lib/firmware/ath10k/QCA6174/hw3.0/board.bin /lib/firmware/ath10k/QCA6174/hw3.0/board.bin.bak
wget https://github.com/FireWalkerX/ath10k-firmware/raw/master/QCA6174/hw3.0/board-2.bin -O /lib/firmware/ath10k/QCA6174/hw3.0/board.bin
reboot now

The Wilocity adapter requires wil6210.fw and wil6210.brd, which don't seem to be widely available online anywhere. I'm not sure what the status is.
 
Reactions: Drazick and IEC

eek2121

Diamond Member
Aug 2, 2005
3,051
4,276
136
I found out the hard way that Windows 10 doesn't like the wifi card that comes with the MSI board. If you start getting BSODs after installing, boot up in safe mode and disable the card.
 
Reactions: lightmanek

eek2121

Diamond Member
Aug 2, 2005
3,051
4,276
136
Guys, there is so much potential with this chip. I am able to undervolt significantly while at the same time increasing clocks. I have yet to find the sweet spot, but wow. If you don't try to squeeze the absolute limit out of the chip as far as clockspeeds are concerned, you will be rewarded by far lower voltages....and therefore, thermals.

EDIT:

So far, for me, the 'sweet spot' appears to be 3.7 GHz @ 1.1V. I can go a bit lower, but 1.075 eventually hard locks after around 30 minutes of stress testing. 1.1V appears to be stable.
 
Last edited:

ajc9988

Senior member
Apr 1, 2015
278
171
116
Guys, there is so much potential with this chip. I am able to undervolt significantly while at the same time increasing clocks. I have yet to find the sweet spot, but wow. If you don't try to squeeze the absolute limit out of the chip as far as clockspeeds are concerned, you will be rewarded by far lower voltages....and therefore, thermals.

EDIT:

So far, for me, the 'sweet spot' appears to be 3.7 GHz @ 1.1V. I can go a bit lower, but 1.075 eventually hard locks after around 30 minutes of stress testing. 1.1V appears to be stable.
What I recommend to everyone is causing the best voltage per multiplier up to the highest that will be used. This is so that if maintenance is needed, or a future bios/uefi or AGESA update makes something unstable, you have a cheat sheet to start from or can down clock until able to perform maintenance.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,752
14,781
136
Guys, there is so much potential with this chip. I am able to undervolt significantly while at the same time increasing clocks. I have yet to find the sweet spot, but wow. If you don't try to squeeze the absolute limit out of the chip as far as clockspeeds are concerned, you will be rewarded by far lower voltages....and therefore, thermals.

EDIT:

So far, for me, the 'sweet spot' appears to be 3.7 GHz @ 1.1V. I can go a bit lower, but 1.075 eventually hard locks after around 30 minutes of stress testing. 1.1V appears to be stable.
The stock vcore of my TR is 1.15, and it does 3.7 ghz all-core turbo all day,.
 
Reactions: nnunn and Drazick

ub4ty

Senior member
Jun 21, 2017
749
898
96
Noctua Threadripper review.. Props to this guy for getting the review out in a timely fashion.. A review that matters to people vs. making repetitive vega rx videos 24/7 hoping to get monstrous view counts for shekels.

Noctua handing non full coverage AIOs their behind.

Noctua NH-U14S is where it's at and is now out of stock at Newegg.

https://www.newegg.com/Product/Prod...octua_threadripper-_-9SIAADY62M6046-_-Product

Got my order in :
Estimated to Ship Out On: 8/28/2017 by End of Day

Also, you guys with Water cooling better make sure those VRMs are cooled. If not, you'd be surprised at how much voltage/extra heat you're going to be running at no matter what you set it to.

Not going to let an AIO touch my baby. Artic freezer 240 is being returned right back to where it came from. So excited !! Got a cooler for my ripper.


> mfw Chaotic42 is going Gentoo w/ a ripper. Wondering what plans he has made
 
Last edited:

eek2121

Diamond Member
Aug 2, 2005
3,051
4,276
136
The stock vcore of my TR is 1.15, and it does 3.7 ghz all-core turbo all day,.
Have you stress tested it? Under Ryzen Master @stock my voltages are all over the place. They don't fix at any one point. The more load that is placed on the CPU, the more voltage is applied and the higher the clocks go. Try firing up Prime95 and opening Ryzen Master and watch what happens.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,752
14,781
136
Have you stress tested it? Under Ryzen Master @stock my voltages are all over the place. They don't fix at any one point. The more load that is placed on the CPU, the more voltage is applied and the higher the clocks go. Try firing up Prime95 and opening Ryzen Master and watch what happens.
well, I ran for days@3.8 using 1.2 vcore, and that was@100% load
 
Reactions: Drazick

ub4ty

Senior member
Jun 21, 2017
749
898
96
Have you stress tested it? Under Ryzen Master @stock my voltages are all over the place. They don't fix at any one point. The more load that is placed on the CPU, the more voltage is applied and the higher the clocks go. Try firing up Prime95 and opening Ryzen Master and watch what happens.
Your VRMs are roasting that's why...
Do you have an AIO cooler or custom water loop? Thus no active air circulation in the VRM area? Do yourself a favor and check your VRM temps
 
Reactions: Drazick
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |