donovan7

I have switched all my x99s to 2680v4, 4 servers, one is a dual x10 SM. Decent core count, decent speed, 40 pci lanes each, and will idle down to around 50W.

5900x cpu Samsung for ssd. Ebay pm981

No. I have the 3in2, the 4in3 and the 5in3... all with the same drives, all run the exact same temps.

Supermicro has some X9 and X10 Motherboards with Dual Processors. You will have enough RAM slots, PCIe sockets... I have one with dual GPUs, dual HBA's, and several Nvme drives with Adapters. Sounds like it would work for you. You wont be able to get there with Ryzen, there are not enough PCIe lanes and sockets. Sever boards are the only way to get there.

Second on the RAM, I have seen this on two of my servers. Both times it was RAM, a single stick went bad. Once I a ebay parts build, another was a stick that was working fine for 3 years and just quit working. Reboot, and from the menu run MEMTEST, Just let it run, let it finished, it might lock up, but you can watch the progress and see about where it is and then you can figure out what stick it was. In my case it would get stuck in MEMTEST around 33gb, so I just counted the ram and found where the stick was where 32GB started in the count. I pulled that stick and replaced, BigPot! done. The other time, I had 8 sticks, I just pulled 4 stick, ran MEMTEST, Passed, so I swapped the other 4 sticks in, ran MEMTEST, it failed, so I then pulled 2 sticks, repeated till I found the single bad stick, done. Easy Peasy, I will say it took several hours each time, the process is long and slow. Hope this helps.

I had this similarly happen last year, it was random about every 3 months. I ended up doing an MEMTEST, and found a single stick of ram went bad, replaced it and the server has been running for 10 months, no issues.

HynesJeff, Check your RAM, run MEMTEST64. Explain; I had a random hardware crash, reboot issue for about 6 months, I looked all over for the issue, but could not figure it out. I ended up running MEMTEST64 and found some bad RAM, it was easy to find out which DIMM it was with isolation. Once I pulled the bad stick out, it was solid for 3 months, RMA'd the ram and all is good again.