Question How do you diagnose silent crashing?

Jul 19, 2022
3
0
10
0
I have a Threadripper 3990x system with 128GB of Ram, Asus board, Running Windows 10 pro etc. It is primarily a render box, rendering V-Ray scenes in 3DS Max. Cooling is handled by an Ice Giant ProSiphon. 1600 watt platinum power supply, yada yada yada.

With some regularity, when all cores are maxed at 100% for more than 5 min, the machine silently crashes. There are no entries into any log files that I can find, and no indication that anything is going wrong. No overclocking, nothing special in configuration of RAM or CPU (just trying to get stability first.) CPU temp stays lower than mid 70's, and the power draw peaks around 600W. So, there is plenty of power and CPU is staying fairly cool. Does anyone have any suggestions of how to diagnose the issue? I would love some verbose logging of EVERYTHING happening at the time of Crash, but there is just nothing. So I keep just trying to eliminate what I think it could be.

I always come here to learn, and have learned so much from this forum. This is actually my first post, as it is such a unique scenario I am struggling to find anyone else with similar issues.

Any help at all is greatly appreciated,

Scott
 

Ralston18

Titan
Moderator
What logs files are you looking at?

Have you looked in Reliability History and Event Viewer?

You can also use Task Manager, Resource Monitor, and Process Explorer to observe what you system is doing or trying to do when crashes occur.

Use all three tools but only one tool at a time.

Process Explorers is free from Microsoft.

https://docs.microsoft.com/en-us/sysinternals/downloads/process-explorer

For the record: make and model PSU? 1600 watts noted. Age? Condition (original to build, new, refurbished, used)?
 
Jul 19, 2022
3
0
10
0
Yes, I have used Reliability, history and Event Viewer. There is really nothing indicating an error prior to the crash, and then warning that the machine was not shut down correctly. No memory dump, no anything from Windows. Nothing in 3DS Max, nothing from V-Ray. I am fairly confident it is a hardware issue because of all of that showing me nothing.

the Power Supply is a EVGA SuperNova 1600 P+, it's about six months old, and was purchased new. I replaced a 1000 watt platinum PSU in this box, as I thought all indications were a lack of power.
 
I have a Threadripper 3990x system with 128GB of Ram, Asus board, Running Windows 10 pro etc. It is primarily a render box, rendering V-Ray scenes in 3DS Max. Cooling is handled by an Ice Giant ProSiphon. 1600 watt platinum power supply, yada yada yada.

With some regularity, when all cores are maxed at 100% for more than 5 min, the machine silently crashes. There are no entries into any log files that I can find, and no indication that anything is going wrong. No overclocking, nothing special in configuration of RAM or CPU (just trying to get stability first.) CPU temp stays lower than mid 70's, and the power draw peaks around 600W. So, there is plenty of power and CPU is staying fairly cool. Does anyone have any suggestions of how to diagnose the issue? I would love some verbose logging of EVERYTHING happening at the time of Crash, but there is just nothing. So I keep just trying to eliminate what I think it could be.

I always come here to learn, and have learned so much from this forum. This is actually my first post, as it is such a unique scenario I am struggling to find anyone else with similar issues.

Any help at all is greatly appreciated,

Scott
Wag.

Proper bios and mobo drivers?

Perhaps boot a copy of memtest86 and let it run for 24 hrs.
 
Jul 19, 2022
3
0
10
0
Follow up: I think this is resolved, I updated the bios and have had no crashes since. I plan to start stress testing it in the coming week or two to see if I can make it crash. So far that does seem to be the answer though. Thanks you both for your input and ideas!
 

ASK THE COMMUNITY