@wendell
Please tell me what ram your using for your system currently 
How is 0d for you ?
This is with the kingston kit but the kingston and gskill kit are about the same.
I can trigger some instability even with qvl kits that result in 00 or 0d errors.
I can pop the gskill in but for me 99.9 mhz and 6000mhz memory with 1800 or 2000mhz fclk is by far thr most stable and hardest to trigger 0d
2 Likes
Thats what i thought. THats exactly what i reported to asus.
It looks like Vcolor is the main one you haven’t gotten your hands on.
Soon ™. I also hsve a 512gb kit on offer from kingston im trying to get to try to see if its the same.
The 128gb /8 dimm kingston fury is fantastic. 6400 and 6000 expo profiles and very stable.
Keep in mind hot vs cold boot can be a huge difference. Letting it be overnight vs let it warm up then hit reset which is kinda lol
1 Like
Lol make sense…damn…i dont like the lower amount of ram but with the price and current state of this newer ddr5 oc memory maybe that kit is the way to go.
I was looking over the asrock board and its outright unsuitable for what i am looking for with my machine. It is in no way like a workstation board. Its like…for oveclocking or just raw compute it would seem.
Very stripped board.
I’m leaning towards that, but more testing is necessary. Right now, it’s like I’m using a whole different machine. There have been absolutely zero stutters or hiccups. I will let the computer idle overnight, in fact, I’ll let it render a project I’ve been slacking on and then idle afterward until I go to work in the morning.
1 Like
What ram are you using again ?
I’m using the G.Skill 128GB 6400 MHz 4x32GB module kit. Profile is set to EXPO I.
So far so good here… was even able to stream and play some Battlebit while listening to music with no hiccups when alt+tabbing or doing things in the background. I have to be honest I was a bit suspicious of the Arc card being the cause of much of the issues I was having, but I didn’t want to believe it. It was to primarily be a nice side AV1 encoder, though I don’t think that’ll happen now.
I’ll revisit these issues at a later date, but for now I will run the system sans Arc card and see what happens. I will leave it running with a load overnight to see what happens.
2 Likes
Have you tried the Kingston 256GB 5600 MT/s kit? Is it less stable than 6000 MT/s?
Ran all night with no issues to report; machine is still super responsive. Temps never got of control with the CPU never breaking beyond 72°C under load. Can’t say for sure about the RAM modules as I’m still getting that absurd temp value.
Once I get everything confirmed to be nice and stable, I would entertain a dedicated loop that would also include the memory modules, but I don’t think I’ll have the space in the Fractal case.
I’m hoping you have some resolution yourself, Kitty!
2 Likes
It is very stable. The 64gb dimms and above get real squrrely. Jedec kits good too.
There seems to be some significsnt board yo board variability and mounting pressure seems to matter a lot.
1 Like
Mounting pressure of the cpu or the heatsink ?
Yes…
The CPU relies on the mounting pressure from the heat sync to help make reliable contact.
When we were working on that 16 node Slurm cluster with a bunch of single socket epyc CPUs mounting pressure was one of the first things we had to rule out before just about any other trouble shooting.
Poor contact can cause memory issues, pcie/device errors, and random crashing or hanging.
This information I am aware of.
What i am asking specifically is which component is causing more of the issues. The threadripper socket locking mechanism or the phsycial heatsink or a combo of both.
@Wendell has said some things about loosening the physical socket but not as much about the actual heatsink.
I had problems loading any OC speed profile for my ram. this system likes to run ram at the stock 2100 speed.
Interesting. What modules are you running? And how many?
2100? Are you running DDR4 RAM?
I am honestly in disbelief at this point. I have replaced every component in this system and replaced my whole entire power grid and this system continues to restart with 0d.
I am currently running the system wwith just 1 gpu and 2 on board nvme connected (1 of the 2 is th system drive) and I am still getting restarts… i do not know were to go from here.Absolutely ridiculous. Threadripper seems cursed for me.