Time Per Frame Problem

haysdb · Post by **haysdb** » Mon Nov 03, 2003 9:39 pm

Electron Microscope reported a "Time Per Frame Problem". I stopped and restarted EM, just in case it was EM having the problem. EM again reported the same problem, so I right-click "Quit" the Folding@Home Windows client, but Task Manager continued to report the app using 100% cpu. I "restarted" F@H, even though it seemed to be running, and then EM reported I was now working on frame 0/100. ANOTHER WU GONE, POOF.

After not too long, EM again started reporting this "Time Per Frame" problem.

I looking in the log file and although it is 11 PM CST on Nov 3, the log file contains the line:

[05:14:22] Working on Unit 05 [November 4 05:14:22]

Why am I seeing times in the log that are 6 hours later than my current time zone? Are the times all reported in GMT?

Here are some of the lines from the log that raise red flags:

Code: Select all

[09:22:47] Preparing to commence simulation
[09:22:47] - Ensuring status. Please wait.
[09:23:07] - Looking at optimizations...
[09:23:07] - Working with standard loops on this execution.
[09:23:07] - Previous termination of core was improper.

[22:27:35] Folding@home Core Shutdown: FILE_IO_ERROR
[22:27:39] CoreStatus = 75 (117)
[22:27:39] Error opening or reading from a file.
[22:27:39] Deleting current work unit & continuing...

[05:13:21] - Checksums don't match (work/wudata_04.goe)
[05:13:22] - Starting from initial work packet

[05:14:04] Folding@home Core Shutdown: FILE_IO_ERROR
[05:14:07] CoreStatus = 75 (117)
[05:14:07] Error opening or reading from a file.
[05:14:07] Deleting current work unit & continuing...

I've gotten the FILE_IO_ERROR at least twice now.

I am completely trashing the F@H directory, downloading a fresh copy, and starting from scratch.

David

Zyzzyx · Post by **Zyzzyx** » Mon Nov 03, 2003 9:55 pm

Yes, the log times are GMT.

As for the other problems... dunno.

TRC-13 · Post by **TRC-13** » Mon Nov 03, 2003 11:03 pm

I had a computer go ape S*@% on me and do the same thing. I ran Quicktech Pro on it and it turned out to be the RAM. But it could be anything.

haysdb · Post by **haysdb** » Tue Nov 04, 2003 1:46 am

I ran across a post on another forum that said this was normal on the first frame, but I was 75% thru a WU when it happened. I suppose I should have just let it run for another hour before I did anything.

I just ran a memory test a couple of days ago, but that doesn't prove it's not a memory problem of course.

David

Post by **NeilBlanchard** » Tue Nov 04, 2003 10:19 am

Hello:

I found that I had to back off my oc frequency by 2mHz (the smallest increment on my mobo) -- it was perfectly stable otherwise, but on about 2 work units, it would freeze. I back off the oc and have not had the problem since...

haysdb · Post by **haysdb** » Tue Nov 04, 2003 11:22 am

NeilBlanchard wrote:Hello:

I found that I had to back off my oc frequency by 2mHz (the smallest increment on my mobo) -- it was perfectly stable otherwise, but on about 2 work units, it would freeze. I back off the oc and have not had the problem since...

I am not overclocked, but I am undervolted - 1.4v vs 1.5v. Might be worth bumping it up .025 just to be a bit safer.

David

haysdb · Post by **haysdb** » Sun Nov 09, 2003 11:22 pm

To answer my own question, the best thing to do when you get this "Time per frame problem" in Electron Microscope is...nothing. It "just happens" once in awhile for no good reason, but will usually work itself out. It happens because EM "expects" a frame to complete in a certain amount of time, and when it doesn't, EM signals this condition. When the frame does complete, EM will go back to being happy again.

David

Time Per Frame Problem

Time Per Frame Problem

are you oc'ing?

Re: are you oc'ing?