ex-Prime95 stops responding, bereft of life

[TheMawn]

As a GIMPS addict, I tend to hover over the P95 tray icon whenever I happen to be around there to see how things are going.

Lately, I've noticed the icon disappearing as soon as I hover over it because Prime95 has just sort of stopped working. I've finally been able to catch it in the act by not minimizing it. (EDIT: That is to say it quietly disappears if it's minimized, but if it isn't, the message pops up and Prime95.exe stays open until I hit "OK") It's the usual Windows blabla.exe has stopped responding message (that you some times get when forcing something to shut down with the task manager).

In the event viewer, under application logs, I find this:

Faulting application name: prime95.exe, version: 28.6.1.0, time stamp: 0x55208535
Faulting module name: prime95.exe, version: 28.6.1.0, time stamp: 0x55208535
Exception code: 0xc0000005
Fault offset: 0x0000000001ce77a1
Faulting process id: 0x1cac
Faulting application start time: 0x01d085d48018ca67
Faulting application path: E:\...\Prime95\prime95.exe
Faulting module path: E:\...\Prime95\prime95.exe
Report Id: 606d686c-f1ed-11e4-bd4b-50465d685f3a


Any idea what this all means?

Lately, I've been screwing around with Excel to solve that stupid IBM Ponder This puzzle, frequently maxing out Excel's memory usage capacity (being a puny 32-bit piece o' crap). However, even at my craziest of three separate instances, I've never gone beyond 3GB total out of my 16GB.

I'm taking a break from that puzzle. For now I'll continue watching for another crash. If there is one, my first order of business will be to get a fresh copy of prime95.exe and all its dependent bits. If it persists, I'll wipe out the most recent save files.

However, I am 95% sure that Excel had... something... to do with it. No idea how, though.

[TObject]

It is an Access Violation. Could be caused by bad memory, or something corrupted or buggy...

[TheMawn]

I still suspect my Excel amblings had something to do with it but now I am more concerned. Less than an hour after my post Prime95 crashed again (so says the event viewer).

I have gotten new copies of everything. Hopefully that fixes things...

[CuriousKit]

What TObject said:

Exception code: 0xc0000005

That hex code means 'access violation', or a 'segmentation fault' for Unix users. As for what's causing it, it could be a program bug, faulty memory or low memory. Assuming bugs have been ironed out over the past 20 years, it's possible that you're running out of memory and Prime95 isn't checking if a request for memory failed (returns a null pointer).

As a user, all I can recommend is running fewer programs, especially background tasks, and seeing if the crash happens as often. Nevertheless, it's interesting that such an exception is raising and hasn't been caught more cleanly by the program.

[axn]

In order of things to try:

1. Reboot
2. Restore to a previous System Restore point.
3. Reformat and reinstall everything.
4. Get a new machine.

[TObject]

If not done already, start with a memory test.

[TheMawn]

Well after a fresh copy of Prime95.exe and all its friends, we've been running fine for 24 hours.

The description of the bug tells me that Excel was being more of a jackass with my system memory than the task Manager said. Something within the program might have gotten messed up too because it took a new copy to get everything healthy again.

Thanks, people. I'll be back if something else goes wrong.

[TheMawn]

There has been two more crashes in the last couple of weeks. I did some digging around in the event viewer and I found something rather strange.

The Security Log with all the Successful or Failed Audits is something I have never really looked through but I noticed that at the exact same second that a Prime95.exe crash is reported, there is a security report like this:

An account was successfully logged on.

Subject:
	Security ID:		SYSTEM
	Account Name:		ANDREW-UBERPC$
	Account Domain:		WORKGROUP
	Logon ID:		0x3e7

Logon Type:			5

New Logon:
	Security ID:		SYSTEM
	Account Name:		SYSTEM
	Account Domain:		NT AUTHORITY
	Logon ID:		0x3e7
	Logon GUID:		{00000000-0000-0000-0000-000000000000}

Process Information:
	Process ID:		0x2b0
	Process Name:		C:\Windows\System32\services.exe

Network Information:
	Workstation Name:	
	Source Network Address:	-
	Source Port:		-

Detailed Authentication Information:
	Logon Process:		Advapi  
	Authentication Package:	Negotiate
	Transited Services:	-
	Package Name (NTLM only):	-
	Key Length:		0

However, I have tons of such reports happening at seemingly random times. So there is such a report at the same time as all prime95 failures, but there isn't a prime95 failure at the same time as each of these reports.

[TheMawn]

I am also seeing a lot of unexpected reboots and a ton of "Corrected Hardware Errors" from the CPU core via the WHEA-logger (whatever the hell that is), despite a significantly reduced core frequency.

I think my i5-3570K is getting sick.

[kladner]

Monitor your voltages, if you don't already. It could be the PSU causing trouble.

[TheMawn]

I have no reduced to stock settings. RAM is also operating at the "default" 1333 MHz which is very sad, but may also provide some insights into whether or not the RAM is the culprit.

To try to determine if the PSU was at fault, I used a fixed voltage instead of an "offset" voltage to see if it fluctuates, and it does not.

RAM is still a suspect. So is the CPU. Motherboard and PSU considerably less so.

Stock settings have the CPU Voltage at a ridiculous 1.35V which is the first thing I will change, if there are no more "corrected hardware errors" from the WHEA-logger and Prime95.exe doesn't fail within 24 hours like it has been doing.