Failing Prime95 torture test with the same error in the same spot (PC Crash) - Page 2

kruoli · 2021-06-11, 01:03

In case of errors while having little load, there is the possibility that the automated energy efficient voltages are now causing problems. I neither know how to test that nor I know if this could a problem in your case.

rapture237 · 2021-06-11, 01:24

Quote:

Originally Posted by kruoli

In case of errors while having little load, there is the possibility that the automated energy efficient voltages are now causing problems. I neither know how to test that nor I know if this is a problem in your case.

That is exactly what I was thinking! It's one of the most plausible causes that I have in mind togheter with ram and vram. A way to test this would be to just use a fixed value for cpu MHz and voltage, and turbo mode off. No clue how to do that with gpu.
Right now my bios settings are default (turbo mode is on by default) with XMP enabled (3200MHz 16-18-18-36 1,35V). By the way I have never overclocked anything, only used intel turbo boost which is a default setting in the bios and it's not overclock.

The hard part of all of this is that I cannot recreate this crash as it doesn't happen under stress and usually it happens once every month, but this week it has happend three times so it might actually help figure out the issue.

I'm a bit tired now and will reconsider this tomorrow, maybe changing some settings in the bios. Again very grateful for your help!

rapture237 · 2021-06-11, 20:45

Quote:

Originally Posted by kruoli

IIRC, the "big" self test of mfaktc might run an hour or a few (the command line option is -st2).

Having no abnormalities at all with Furmark might be a hint, but of course I can't say this for sure. When done with my suggested stress tests and still not having occured a specific problem, I'm interested in what others might suggest.

So I've been suggested in addition to furmark to test with 3rd Mark, here are the results after more than 10 cycles of test. On the hwinfo64 screen 3rd column there's all gpu sensors including VRM and a new sensor added a few months ago "gpu hot spot temperature". Here's an article about it https://www.tomshardware.com/news/nv...spot-detection

Also I've tried mfaktc but the program won't start at all, maybe because I have an nvidia gpu?

kruoli · 2021-06-14, 11:07

Quote:

Originally Posted by rapture237

Also I've tried mfaktc but the program won't start at all, maybe because I have an nvidia gpu?

This program is optimised for nVidia, so it should work. Have you started it from the command line? If so, what error is displayed? If not, please start it from the command line and attach the -st2 parameter.

rapture237 · 2021-06-14, 11:52

Hey kruoli, I haven't started it from the command line, it was late night and didn't have much time to read, will try again, thanks.

Anyway an update, I'm currently testing the system with an old Kingston module 8gb ddr4 2133MHz CL15 https://www.kingston.com/datasheets/kvr21n15s8_8.pdf we'll see if anything happens. Also all the gpu stress tests I ran didn't crash.

rapture237 · 2021-08-11, 12:06

Hello everyone, I've been testing the system with this old ddr3 module for two months now and I've had no crashes. Today I've received and installed two new ddr3 modules from corsair rma, I'm currently running default bios settings and XMP I, we'll see how it goes.

kriesel · 2021-08-11, 15:09

Quote:

Originally Posted by rapture237

I haven't started it from the command line, ..., will try again

Mfaktc getting started guide & other info is here

rapture237 · 2021-08-11, 16:35

Hi, thanks for the info. I have run and passed multiple gpu tests and the system hasn't crashed for 2 months straight with different ram so I guess the problem is circumscribed to that. Will keep this in mind for the future.

2021-06-11, 01:03	#12
kruoli "Oliver" Sep 2017 Porta Westfalica, DE 23·71 Posts	In case of errors while having little load, there is the possibility that the automated energy efficient voltages are now causing problems. I neither know how to test that nor I know if this could a problem in your case. Last fiddled with by kruoli on 2021-06-11 at 01:30 Reason: Conjugative. Scrap that. Apparently it's called subjunctive.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Prime95 v28.5.1 Error at Torture test start	RMSe17	Information & Answers	2	2014-12-16 18:55
Athlon 64 X2 5600+ Failing Torture Test	lbthe3rd	Information & Answers	29	2007-08-05 02:05
torture test failing in 1 minute	ghatothkach	Hardware	5	2005-01-14 09:50
Failing Torture Test..	jugbugs	Hardware	12	2004-03-25 02:37

2021-06-14, 11:52	#16
rapture237 Jun 2021 2²×3 Posts	Hey kruoli, I haven't started it from the command line, it was late night and didn't have much time to read, will try again, thanks. Anyway an update, I'm currently testing the system with an old Kingston module 8gb ddr4 2133MHz CL15 https://www.kingston.com/datasheets/kvr21n15s8_8.pdf we'll see if anything happens. Also all the gpu stress tests I ran didn't crash.

2021-08-11, 12:06	#17
rapture237 Jun 2021 2²×3 Posts	Hello everyone, I've been testing the system with this old ddr3 module for two months now and I've had no crashes. Today I've received and installed two new ddr3 modules from corsair rma, I'm currently running default bios settings and XMP I, we'll see how it goes.

2021-08-11, 16:35	#19
rapture237 Jun 2021 1100₂ Posts	Hi, thanks for the info. I have run and passed multiple gpu tests and the system hasn't crashed for 2 months straight with different ram so I guess the problem is circumscribed to that. Will keep this in mind for the future.