Threadripper BUILDERS thread

[eek2121]

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

This is an AiO cooler and all temps are normal. There is a TON of cool air flowing into the case.

EDIT: To be more specific, when under load, VRM temps are in the mid 50s when the system is under load.

It looks like a bug in Ryzen Master. HWInfo shows a completely different situation. According to Ryzen Master things are bouncing around 1.4 V at idle. According to HWInfo, it's actually at 1.144V currently, and the max is 1.144.

Edit #2: With no changes to Ryzen master (reset to defaults), clockspeed is also hovering around 3.7 GHz at idle. Very odd.

 

[ub4ty]

This is an AiO cooler and all temps are normal. There is a TON of cool air flowing into the case.

EDIT: To be more specific, when under load, VRM temps are in the mid 50s when the system is under load.

It looks like a bug in Ryzen Master. HWInfo shows a completely different situation. According to Ryzen Master things are bouncing around 1.4 V at idle. According to HWInfo, it's actually at 1.144V currently, and the max is 1.144.

Ah', in that case (two software packages reporting different info than it's just that. One of them is wrong. That being said, if you set Voltage limits, that sets what the processor will demand. If your vrms aren't delivering it, it will push them harder thus how voltage can fluctuate even after you've set them to certain limits. Think of it as an effective value that pushes other components to meet it.

 

[Markfw]

My AIO radiator is right above the vrm. full load@3.9 IS 45-47C on vrm (I think) PER HWInfo Here:

 

[ub4ty]

My AIO radiator is right above the vrm. full load@3.9 IS 45-47C on vrm (I think) PER HWInfo Here:

Very nice #'s. You really won the silicon lottery being able to run 3.9 @ 1.21/1.25v. I'm pretty sure that's not a common # or possibility. This is likely why your AIO is able to handle things within the temp range that I see. 1.35V and you'd be toast. I'm not counting on this being a regular voltage at those clocks thus why current AIOs are out of the question. Btw, is this with 3600 CL16 ram settings?

 

[Markfw]

Very nice #'s. You really won the silicon lottery being able to run 3.9 @ 1.21/1.25v. I'm pretty sure that's not a common # or possibility. This is likely why your AIO is able to handle things within the temp range that I see. 1.35V and you'd be toast. I'm not counting on this being a regular voltage at those clocks thus why current AIOs are out of the question. Btw, is this with 3600 CL16 ram settings?

Actually most of the TR chips appear to run like this. IEC's box does 4 ghz@ the same vcore I am working at 3.9., and yes, thats 3600 CL16 @ rated timings

 

[Khanan]

@Markfw : what do you plan to do with the build if I may ask? Curious.

 

[Markfw]

@Markfw : what do you plan to do with the build if I may ask? Curious.

Simple... As everyone else knows here (you are new I see) I do distributed computing. In my specific case almost all Folding@home. I am number 36 in the world, and soon to gain another place or two.

 

[Khanan]

Simple... As everyone else knows here (you are new I see) I do distributed computing. In my specific case almost all Folding@home. I am number 36 in the world, and soon to gain another place or two.

That's great, good luck with that.

 

[ddogg]

Just reserved an open box 1950x at my microcenter branch for 899 + tax. Comes with a 15 day microcenter return and regular manufacturer warranty. Not sure why it was returned, but I think it's worth checking it out at that price. Will be picking it up tomorrow.

 

[VirtualLarry]

If I were in the market, a $100-off-of-$1000 open-box CPU would be tempting. But I wonder too, why it might have been returned?

 

[ddogg]

If I were in the market, a $100-off-of-$1000 open-box CPU would be tempting. But I wonder too, why it might have been returned?

I have a 15 day return period on it so it should be enough time for me to put it through it's paces and find out. Maybe it didn't OC as well or he didn't like the platform. For 10% off, it's worth trying IMO.

 

[eek2121]

I was finally able to setup push/pull today using a case fan and an old Rosewill fan from long ago. I didn't want to blow any more money on this build unless necessary. The wife wants me to spend some on her

Push/Pull is the way to go for AIOs. It handles temps much better under load. I can only imagine how much more effective a 360mm rad is as opposed to my 240. Temps steady in the mid 60s @ 3.9GHz.

 

[SangamonTaylor]

But I wonder too, why it might have been returned?

It won't fit in his socket 939 board.

 

[IEC]

If I were in the market, a $100-off-of-$1000 open-box CPU would be tempting. But I wonder too, why it might have been returned?

Might be one of the people who got a Asus X399 board that has the marginal/defective Foxconn socket retention bracket. And so gave up on TR.

 

[Chaotic42]

Now that my hatred for this computer is gone, I can tell my full story

I found out that I had about $6,000 more than I thought I did. I broke my lease in Cincinnati early and was responsible for six more months of rent. It turns out that they leased it almost immediately, so I got my last month of rent back, plus my deposit, plus I don't have to pay them anymore. Filled with joy, I drove over to Microcenter and picked up a TR 1950X, an Asus Zenith Extreme X399 board, 32GB of memory, an AIO cooler, a Corsair AIR540 case (which I love, by the way), and an M.2 drive. I got home, went to sleep, and started putting the computer together the next day. That is when sadness struck...

First, my CPU had a damaged orange bracket. Half of the little clips on it were broken, and it was slightly cracked at one end. I gently put it into the mounting slot and closed it. Then Foxconn struck. A ton of people have had issues with the socket retention bracket not closing well enough to be able to screw in the star screw labelled 1. I tried wiggling it, shifting it, nothing. In attempts to figure out the problem, I removed the CPU cassette several times. One time, as I was holding it, the CPU fell out of the orange bracket right onto the socket. It seemed to be in there correct and I had no easy way to remove it since there was no orange bracket.

Hoping against hope, I went forward assuming that the orange bracket wasn't necessary for clearance or stability purposes. Putting as much of my weight as I could (seriously, I was scared) onto the pointy screwdriver provided by AMD, I finally got the star screws to bite.

I set about building the system only to find that my video card didn't work. I wasn't sure what the problem was exactly because the helpful LED display on the Asus motherboard has some sort of problem and I can only read part of it. Swapping out video cards, I finally got a POST. Holy crap. I flashed the BIOS and set about installing Linux Mint (Ubuntu was taking hours to download).

Of course, being a new motherboard, the wifi drivers weren't easy to come by. After some digging, I found the solution I posted above - essentially one of the adapters isn't supported yet and the other has corrupted firmware in the main repositories. Finally, way too long after I started, I have a computer.

At the time, if I had super powers, I'd have flung AMD, Asus, and Foxconn into the sun. Now, I'm not sure how I feel. It definitely feels like Asus didn't do proper QC since so many people are having this bracket problem and I'm having the LED display problem. For $570, I want more quality and fewer stupid stickers. Anyway, I've got POV-Ray installed and I'll be taking the system out for some tests this weekend. If anyone has any questions, feel free to hit me up. I don't regret the build, but I'm just disappointed in the quality issues I've encountered along the way. Maybe I'm too needy.

 

[VirtualLarry]

I don't regret the build, but I'm just disappointed in the quality issues I've encountered along the way. Maybe I'm too needy.

Sorry to hear about your troubles. That's what we call "bleeding edge". (Especially if you cut yourself during the build.)

I didn't have too many issues to speak of with my Ryzen 5 1600 builds, although, I transplanted the CPU/mobo/RAM/VGA from one set of cases to another (5 expansion slots, so I could put two video cards in), and in the process, I re-mounted the CPU. I think that I used MX-4 paste. Well, sadly, it was like glue, much like the OEM paste that shipped on AM2/AM2+/AM3 heatpipe heatsinks, that after a few months of being applied, becomes like glue.

I twisted the stock heatsink (after unscrewing it from the backplate), but it wouldn't go free. So I pulled on it, and the whole dang CPU came out of the ZIF socket with the heatsink! I had heard of this happening with AM2/AM3, but not with Ryzen. Scary stuff. Thankfully, all of the pins came out intact too. It was a PITA to get that CPU free though.

(I had initially made a mistake mounting the stock cooler, and the "AMD" nub on one side, I put towards the RAM slots, when it gets in the way if you do it that way.)

And I just had to take the PC apart again, because one of the fans was rubbing against a wire (the rear exhaust fan, on its own cable). (I really should learn some cable-management skills some day.)

 

[hasu]

Now that my hatred for this computer is gone, I can tell my full story

I found out that I had about $6,000 more than I thought I did. I broke my lease in Cincinnati early and was responsible for six more months of rent. It turns out that they leased it almost immediately, so I got my last month of rent back, plus my deposit, plus I don't have to pay them anymore. Filled with joy, I drove over to Microcenter and picked up a TR 1950X, an Asus Zenith Extreme X399 board, 32GB of memory, an AIO cooler, a Corsair AIR540 case (which I love, by the way), and an M.2 drive. I got home, went to sleep, and started putting the computer together the next day. That is when sadness struck...

First, my CPU had a damaged orange bracket. Half of the little clips on it were broken, and it was slightly cracked at one end. I gently put it into the mounting slot and closed it. Then Foxconn struck. A ton of people have had issues with the socket retention bracket not closing well enough to be able to screw in the star screw labelled 1. I tried wiggling it, shifting it, nothing. In attempts to figure out the problem, I removed the CPU cassette several times. One time, as I was holding it, the CPU fell out of the orange bracket right onto the socket. It seemed to be in there correct and I had no easy way to remove it since there was no orange bracket.

Hoping against hope, I went forward assuming that the orange bracket wasn't necessary for clearance or stability purposes. Putting as much of my weight as I could (seriously, I was scared) onto the pointy screwdriver provided by AMD, I finally got the star screws to bite.

I set about building the system only to find that my video card didn't work. I wasn't sure what the problem was exactly because the helpful LED display on the Asus motherboard has some sort of problem and I can only read part of it. Swapping out video cards, I finally got a POST. Holy crap. I flashed the BIOS and set about installing Linux Mint (Ubuntu was taking hours to download).

Of course, being a new motherboard, the wifi drivers weren't easy to come by. After some digging, I found the solution I posted above - essentially one of the adapters isn't supported yet and the other has corrupted firmware in the main repositories. Finally, way too long after I started, I have a computer.

At the time, if I had super powers, I'd have flung AMD, Asus, and Foxconn into the sun. Now, I'm not sure how I feel. It definitely feels like Asus didn't do proper QC since so many people are having this bracket problem and I'm having the LED display problem. For $570, I want more quality and fewer stupid stickers. Anyway, I've got POV-Ray installed and I'll be taking the system out for some tests this weekend. If anyone has any questions, feel free to hit me up. I don't regret the build, but I'm just disappointed in the quality issues I've encountered along the way. Maybe I'm too needy.

Very funny post! I had similar kind of problems when I first started on my bike. Lesson learned - nothing wrong in using a little hand-holding when it is really needed.

 

[jmelgaard]

Simple... As everyone else knows here (you are new I see) I do distributed computing. In my specific case almost all Folding@home. I am number 36 in the world, and soon to gain another place or two.

Just to be nitpicking... So what you then mean is that you kindly lend your hardware to other people that does distributed computing?... Or are you actually also part of developing the software and algorithms behind folding@home?...

 

[Markfw]

Just to be nitpicking... So what you then mean is that you kindly lend your hardware to other people that does distributed computing?... Or are you actually also part of developing the software and algorithms behind folding@home?...

Stanford has a program. You download their software, run it at your house, and it computes the units, and sends the results back.

 

[jmelgaard]

Stanford has a program. You download their software, run it at your house, and it computes the units, and sends the results back.[/QUOTE

I know what folding@home is... but you didn't answer my question... Although I think I can take it as an answer that you don't actually do Distributed computing your self, your just lending your hardware to others that do.

 

[IEC]

I know what folding@home is... but you didn't answer my question... Although I think I can take it as an answer that you don't actually do Distributed computing your self, your just lending your hardware to others that do.

Why the semantics? The whole concept of distributed computing is to divide the work of a project out into work units that can then be processed and validated by multiple users/machines. Thus he is directly participating in a DC project. As he has full control over his machines and whether they accept work or not, he is not "lending" his hardware to others. Not to mention he receives no compensation for doing so.

 

[jmelgaard]

Why the semantics? The whole concept of distributed computing is to divide the work of a project out into work units that can then be processed and validated by multiple users/machines. Thus he is directly participating in a DC project. As he has full control over his machines and whether they accept work or not, he is not "lending" his hardware to others. Not to mention he receives no compensation for doing so.

Because of interest, I am a software architect so it would be interesting if the case was the later...

The whole concept of things like folding@home, seti@home is that you lend your hardware to do calculations for others, the fact that he has full control doesn't make any difference, this was quite unique when it first arrived because before you where use to do such things in big super-computers build in large data centers... The fact that he is in full control doesn't change that fact, it just means he sets the terms for the lend...

The mere concept of building such software and infrastructure is super interesting...
It's not to belittle what he does, it just not as interesting to me (professionally)...

 

[Markfw]

Because of interest, I am a software architect so it would be interesting if the case was the later...

The whole concept of things like folding@home, seti@home is that you lend your hardware to do calculations for others, the fact that he has full control doesn't make any difference, this was quite unique when it first arrived because before you where use to do such things in big super-computers build in large data centers... The fact that he is in full control doesn't change that fact, it just means he sets the terms for the lend...

The mere concept of building such software and infrastructure is super interesting...
It's not to belittle what he does, it just not as interesting to me (professionally)...

Well, technically that is also incorrect. Example:

Normal day, I wake up, and F@H has been running all night on the 2 1080TI cards all night, and the threadripper cores have been running WCG (another DC project). I decide to play a game, so I pause the GPU's on F@H, and change the BOINC configuration for WCG to only use 70% of the CPU's, then I play for a few hours. When done, I set everything back to the way it was in the morning.

My point is that I DONATE computer resources at differing levels to DC projects to help cure cancer

 

[ddogg]

Might be one of the people who got a Asus X399 board that has the marginal/defective Foxconn socket retention bracket. And so gave up on TR.

You are actually very close lol. I went in today and picked it up and asked the guy why it was returned. He picked up the zenith + TR and returned everything back because he decided he didn't need as much CPU. Weird but his loss, my win

 

[jmelgaard]

My point is that I DONATE computer resources at differing levels to DC projects to help cure cancer

Well that would have the exact same meaning to me, but if you like that wording better then sure... >.<...
As I said, it wasn't to belittle what you did... It just didn't have the same professional interest to me....