I MIGHT HAVE UNDERESTIMATED....
- Thread starter oddssodds
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
- the first board had a dead socket #2 (both cpu's worked in socket #1)
- The second board got the water fountain treatment
- The third board got the idiot with the air hose treatment
- The fourth board is installed. but cpu2 not working. This time it isn't because of a bad socket. Only one cpu works in socket #1. The cpu that got the water treatment doesn't work in socket #1 anymore. Acts the same as first board, but different cause.
yes, that was the board with bent socket 2 pins. But both cpu's worked in socket one by themselves. This time cpu (water) does not work in socket 1
So after a first start up water leak disaster, claiming the motherboard and one of the processors, everything is replaced and all the technical details have been sorted.
The outstanding issues are backing down Idrac preboot tests, to lessen start up time and unrestrict processor safety a bit.
Build as follows:
Runs mostly in the 30c range. Extreme load seems to put stress on certain cores 60c to 80c
Happy to be done. Thx to @RolloZ170 and others who helped
The outstanding issues are backing down Idrac preboot tests, to lessen start up time and unrestrict processor safety a bit.
Build as follows:
- Thermaltake X500 case (heavily modified)
- R740 Dell Mobo (vrm upgraded)
- 192 gb of Crucial RDIMM 2933
- x2 Xeon 8252c
- EK Annihilator Pro water blocks
- Alphacool Eisbecher Aurora D5 res/pump
- Alphacool NexXxos XT45 Full Copper Radiator, 420mm
- Thermalrite TL-B14015 EXTREM push pull fan set up
- Noctua 60 mm vrm fans with copper heatsink
- USB 3.2 Gen 1 Motherboard Header 19 Pin Internal USB Hub Controller-Type-A to 19-Pin + Type-E (USB 3.2 Front Panel Header)
- OCCTO fan header and 6 fan spoof circuit boards
- Dell poweredge front cp cable and modified interface (runs OCCTO and Clover boot)
- Dell riser 1a PCIe two x16 slots
- Dell NVMe caddy with room for two additional drives (riser 1)
- Nvidia 3090ti (riser 1)
- Dell riser 2a (modified)
- ST113 PCIe Sound Card, Dual-Chip Audio Decoder with 3.5mm MIC-in, 3.5mm PH-Out, SPK-Out, SPDIF Optical and Coaxial Digital Output (riser 2)
- Samsung 980 pro (NVMe caddy)
- Dell 2000w power supply
Runs mostly in the 30c range. Extreme load seems to put stress on certain cores 60c to 80c
Happy to be done. Thx to @RolloZ170 and others who helped
Attachments
Not quite yet!
So, if you look at the core temps on cpu1, you can see that four of the cores were running 20 - 25c higher than the others.
So I pulled the processor and replaced the thermal pad
that fixed it, but I have another issue.
because the right front control panel has been repurposed it sits here
I have been getting critical errors on inlet temp, which I have been ignoring, 38-39c.
Now I am getting sever throttling. My first stress tests were all at x45. The last were at x15
It turns out that the inlet temp sensor is on that front cp. It was really hot yesterday and the coolest I could get it was 36c, but only doing this:
There is no room for a fan on that side of the case.... maybe a vrm fan, but just maybe. That's a bit of a house of cards.
I went through Idrac with ai, trying to see if I could just raise the critical threshold, but no success as of yet.
Anyone have any ideas?
@RolloZ170
So, if you look at the core temps on cpu1, you can see that four of the cores were running 20 - 25c higher than the others.
So I pulled the processor and replaced the thermal pad
that fixed it, but I have another issue.
because the right front control panel has been repurposed it sits here
I have been getting critical errors on inlet temp, which I have been ignoring, 38-39c.
Now I am getting sever throttling. My first stress tests were all at x45. The last were at x15
It turns out that the inlet temp sensor is on that front cp. It was really hot yesterday and the coolest I could get it was 36c, but only doing this:
There is no room for a fan on that side of the case.... maybe a vrm fan, but just maybe. That's a bit of a house of cards.
I went through Idrac with ai, trying to see if I could just raise the critical threshold, but no success as of yet.
Anyone have any ideas?
@RolloZ170
when I cool the right ear, the problem goes away. Hoping to solve it without putting stupid fans on the case. The temp on that sensor is 32 right now. This is what the cinebench r23 looks like on hwinfo
It's just that sensor and if I could raise the threshold to 40c, all problems gone
So I gave up and just put a fan in it. also modded the intrusion switch and got rid of the "your server is open"
Ran another cinebench to check if the clocks were being throttled. The same cores 1,4,8,11 were hitting in the mid to upper 90's. So it looks like the wrong cpu got the water. Anyway, fine for now. I'll change it out at some point.
Ran another cinebench to check if the clocks were being throttled. The same cores 1,4,8,11 were hitting in the mid to upper 90's. So it looks like the wrong cpu got the water. Anyway, fine for now. I'll change it out at some point.
So the fan fixed the issue and I sorted the case open issue, BUT...
Still have the same 4 cores going 20c higher than all the others. Cpu1 cores 1 , 4 , 8 , 11
I reseated it. Then I took it out and changed the phase change thermal pad. Nothing makes a difference. Obviously the cooling is adequate. Is there some sort of BIOS setting I am missing? Is it another shit processor?
Still have the same 4 cores going 20c higher than all the others. Cpu1 cores 1 , 4 , 8 , 11
I reseated it. Then I took it out and changed the phase change thermal pad. Nothing makes a difference. Obviously the cooling is adequate. Is there some sort of BIOS setting I am missing? Is it another shit processor?
those OEM 240W run over spec. they last 3 years only, but nobody expexted them to live longer.
i had similar with P-8124 (B1 240W OEM)
TIM was dried out, delid and reTIM solved the issue.
Yea, tried that. didn't matter. I will say that the replacement for the water damaged cpu looked a lot newer.
Oh you mean the heat spreader. That never occurred to me.
I’ve seen it done with a vice. How hard is it to do?
Reseating the CPU twice and changing the thermal paste between the water block and the CPU pretty much indicates that it is the CPU, given the exact same cores are overheating.
It wouldn’t be the motherboard? The pins look really good.
I was very close to ordering another CPU from the guy who sent the last one. It looked a lot newer than the one I took out all the capacitors were shiny and it looked like it had almost no usage.
With that said, I can’t keep spending money on this…. I have really been dragged and it’s attractive to fix.
thx @RolloZ170