Fatal error: NaN detected

SG Distributed Computing Teams discussion, S.E.T.I., Folding@Home, as well as other types of cloud/distributed computing, AI, machine learning, cryptocurrency mining.
Post Reply
User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Fatal error: NaN detected

Post by goobee »

I've been getting a lot of these NaN errors. My Opty box is not overclocked and passes 4 instances of Prime95 torture test just fine.

I removed the -advmethods so that I don't get any beta units and it sorta helped. I removed -forceasm and it made no difference.

Any idea what's wrong?

Box consists of:

2 x Opty 270
2 x Thermalright XP90 hsf
4 x 1gbs PC3200 ecc/reg ram (all single rank)
Asus K8N-DL mobo (tried 2 boards, same errors)
Generic X700 PCIe video Card
500watts Antec EPS powersupply
Windows XP x64

4 x instances of F@H
"Signature?........We don't need no stinkin' Signature!!!"
User avatar
YARDofSTUF
Posts: 70006
Joined: Sat Nov 11, 2000 12:00 am
Location: USA

Post by YARDofSTUF »

User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Post by goobee »

Hi YoS,

I think I'm OK now. After I purged my WUs, all 4 instances have been running fine for the past 6 hours or so. Before, it wouldn't even do an hour before crashing and burning. I'm leaving the -advmethods off, I think this flag let's beta units download. It must be bad beta units that are unstable. In any case, I put -forceasm back on and it's still going OK.
"Signature?........We don't need no stinkin' Signature!!!"
User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Post by goobee »

Nope, I was wrong. Ran fine for like 12 hours but still errored out after a while. :(
"Signature?........We don't need no stinkin' Signature!!!"
User avatar
rickoic
Senior Member
Posts: 2172
Joined: Sat Aug 31, 2002 12:00 pm
Location: North Mississippi

Post by rickoic »

NaN stands for "Not A Number".

Means that in a calculation that expected a number to be passed to it got a alpha-numeric character instead.

Believe I read someplace before that this could be a RAM problem.

You might want to run a memory test on your ram just to check it out.

Tks
Rick
Prostate Cancer survivor, too late for me with a cure, but maybe I can help save others just like me. Wife brest cancer. Cured.
User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Post by goobee »

Hi Rick,

I'm running memtest86 right now. So far, no errors. I'll leave it going over night and see what happens.
"Signature?........We don't need no stinkin' Signature!!!"
User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Post by goobee »

Almost 10 hours, no errors found thus far. Still running.
"Signature?........We don't need no stinkin' Signature!!!"
User avatar
YARDofSTUF
Posts: 70006
Joined: Sat Nov 11, 2000 12:00 am
Location: USA

Post by YARDofSTUF »

probably not memory then.

Its not the in C:\program files\ folder is it?
User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Post by goobee »

YARDofSTUF wrote:probably not memory then.

Its not the in C:\program files\ folder is it?
No, I have them on the "D" Drive as such:

Main Directory

D:\Folding

Sub Folders

D:\Folder 1
D:\Folder 2
D:\Folder 3
D:\Folder 4

This is how I have been folding for years now.

No memory errors detected after 13 hours so I'm pretty sure the ram is OK.

So far I've swapped mobos, swapped power supplies and CPUs. The memory is ruled out as a result of my testing. The video card is the only thing left I haven't considered and I doubt it can be screwing with F@H. I guess the only thing left is to do a new Windoze installation at this point.

Oh, one last thing. This box is my stand alone MySQL server. All it does is serve MySQL and Fold 24/7.
"Signature?........We don't need no stinkin' Signature!!!"
User avatar
goobee
Advanced Member
Posts: 772
Joined: Sat Apr 13, 2002 12:55 am
Location: Sunny California

Post by goobee »

OK, it's fixed. I had a bad Windows x64 installation it appears. I reinstalled everything from scratch last night and after 18 hours, no F@H errors.

Talk about a whole bunch of work for nothing. :irate:
"Signature?........We don't need no stinkin' Signature!!!"
Post Reply