SMP: A1 core not so stable anymore?

Come join the... uh... er... fold.

Moderators: just brew it!, farmpuma

SMP: A1 core not so stable anymore?

Postposted on Sat Nov 01, 2008 8:56 am

I have been having problems with project 2665 with the A2 core so what I did was to just keep fishing for 2605's whenever I see *_a1.exe downloaded. Looks like they have moved 2665 to the A1 core now and it seems stability has not improved for me. Same old hang issue. :-?

Anybody notice any other changes?

Note: I am using notfred's diskless cd in a 64-bit Linux VM hosted by 32-bit XP.
Image
The Model M is not for the faint of heart. You either like them or hate them.

Gerbils unite! Fold for UnitedGerbilNation, team 2630.
Flying Fox
Gerbil God
 
Posts: 24297
Joined: Mon May 24, 2004 2:19 am

Re: SMP: A1 core not so stable anymore?

Postposted on Sun Nov 02, 2008 8:58 am

Which hang exactly are you seeing? The core_a1's used to do "FINISHED_UNIT" in the logfile and not follow it with "CoreStatus" and that needs a kill -9 to the hung core. The core_a2's used to do "Number of Units Completed" without "Preparing to get new work unit" in the logfile and that needs a kill -CONT to the hung core. If you are running the latest version of my diskless folding stuff (it should show the version on the screen / webpage and you can compare with the download website), it should check and catch both of those and do the appropriate kill command to clear the hang. Can you post the end of the FAHlog.txt and also the hanglog.txt if it is not catching these hangs and fixing them?
notfred
Grand Gerbil Poohbah
 
Posts: 3716
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada

Re: SMP: A1 core not so stable anymore?

Postposted on Sun Nov 02, 2008 3:03 pm

I shut down the VM for a while hoping I'll fish some 2605's later. So I lost the log files.

Basically it just happens in the middle of folding, 10% or 90% does not matter. It just seems to "hang" and do nothing, CPU % drops to minimal instead of the usual 80%+. Restarting won't help and it will eventually go past the deadline, making the current WU worthless anyway. So i have resorted to just trashing the 2665's and fish for the 2605. I know, I already said it is bad for the science but I can't seem to ever return a good 266x WU. :x

I'm pretty sure it's not your stuff notfred. Just stupid SMP instability. I'll grab the latest folding CD during this downtime and post more info later if you like.
Image
The Model M is not for the faint of heart. You either like them or hate them.

Gerbils unite! Fold for UnitedGerbilNation, team 2630.
Flying Fox
Gerbil God
 
Posts: 24297
Joined: Mon May 24, 2004 2:19 am

Re: SMP: A1 core not so stable anymore?

Postposted on Sun Nov 02, 2008 7:19 pm

OK, that sounds more like a core crashing and not being picked up. My stuff can't pick that up, although if anyone can suggest a way I'm open to trying it.
notfred
Grand Gerbil Poohbah
 
Posts: 3716
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada


Return to TR Distributed Computing Effort

Who is online

Users browsing this forum: No registered users and 4 guests