Page 1 of 1

Update_Core status on a work unit.

Posted: Mon May 04, 2020 1:40 pm
by scm2000
I've got a work unit stuck in Update_Core:

-----------------------------------------

Code: Select all

12:45:59:WU00:FS01:FahCore 0x22 started
12:45:59:WU00:FS01:0x22:*********************** Log Started 2020-05-04T12:45:59Z ***********************
12:45:59:WU00:FS01:0x22:*************************** Core22 Folding@home Core ***************************
12:45:59:WU00:FS01:0x22:       Type: 0x22
12:45:59:WU00:FS01:0x22:       Core: Core22
12:45:59:WU00:FS01:0x22:    Website: https://foldingathome.org/
12:45:59:WU00:FS01:0x22:  Copyright: (c) 2009-2018 foldingathome.org
12:45:59:WU00:FS01:0x22:     Author: John Chodera <john.chodera@choderalab.org> and Rafal Wiewiora
12:45:59:WU00:FS01:0x22:             <rafal.wiewiora@choderalab.org>
12:45:59:WU00:FS01:0x22:       Args: -dir 00 -suffix 01 -version 706 -lifeline 9736 -checkpoint 15
12:45:59:WU00:FS01:0x22:             -gpu-vendor nvidia -opencl-platform 1 -opencl-device 2 -cuda-device
12:45:59:WU00:FS01:0x22:             2 -gpu 2
12:45:59:WU00:FS01:0x22:     Config: <none>
12:45:59:WU00:FS01:0x22:************************************ Build *************************************
12:45:59:WU00:FS01:0x22:    Version: 0.0.2
12:45:59:WU00:FS01:0x22:       Date: Dec 6 2019
12:45:59:WU00:FS01:0x22:       Time: 21:30:31
12:45:59:WU00:FS01:0x22: Repository: Git
12:45:59:WU00:FS01:0x22:   Revision: abeb39247cc72df5af0f63723edafadb23d5dfbe
12:45:59:WU00:FS01:0x22:     Branch: HEAD
12:45:59:WU00:FS01:0x22:   Compiler: Visual C++ 2008
12:45:59:WU00:FS01:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
12:45:59:WU00:FS01:0x22:   Platform: win32 10
12:45:59:WU00:FS01:0x22:       Bits: 64
12:45:59:WU00:FS01:0x22:       Mode: Release
12:45:59:WU00:FS01:0x22:************************************ System ************************************
12:45:59:WU00:FS01:0x22:        CPU: Intel(R) Celeron(R) CPU G3930 @ 2.90GHz
12:45:59:WU00:FS01:0x22:     CPU ID: GenuineIntel Family 6 Model 158 Stepping 9
12:45:59:WU00:FS01:0x22:       CPUs: 2
12:45:59:WU00:FS01:0x22:     Memory: 7.70GiB
12:45:59:WU00:FS01:0x22:Free Memory: 3.52GiB
12:45:59:WU00:FS01:0x22:    Threads: WINDOWS_THREADS
12:45:59:WU00:FS01:0x22: OS Version: 6.2
12:45:59:WU00:FS01:0x22:Has Battery: false
12:45:59:WU00:FS01:0x22: On Battery: false
12:45:59:WU00:FS01:0x22: UTC Offset: -4
12:45:59:WU00:FS01:0x22:        PID: 8464
12:45:59:WU00:FS01:0x22:        CWD: C:\Users\steph\AppData\Roaming\FAHClient\work
12:45:59:WU00:FS01:0x22:         OS: Windows 10 Pro
12:45:59:WU00:FS01:0x22:    OS Arch: AMD64
12:45:59:WU00:FS01:0x22:********************************************************************************
12:45:59:WU00:FS01:0x22:Project: 13402 (Run 33, Clone 183, Gen 0)
12:45:59:WU00:FS01:0x22:Unit: 0x0000000112bc7d9a5eafa8f54d8ddd97
12:45:59:WU00:FS01:0x22:ERROR:110: Need version 0.0.5
12:45:59:WU00:FS01:0x22:Folding@home Core Shutdown: CORE_OUTDATED
12:46:00:WARNING:WU00:FS01:FahCore returned: CORE_OUTDATED (110 = 0x6e)
What should I do?

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 1:49 pm
by JimboPalmer
So far as I know, those are in beta. (i am not a member of the beta team)

Are you a member of the beta team? This should be posted there.

Do you have a beta flag even though you are not a member of the beta team? Remove the beta flag.

Are you getting this as a regular volunteer? This is the place to raise the issue.

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 1:50 pm
by scm2000
I'm not on the beta team... I never set any beta flag.. this is just a stock install of FAH.

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 1:54 pm
by JimboPalmer
I see this in advanced, maybe it leaked out.

add to this thread

viewtopic.php?f=24&t=35063&p=332176&hilit=13400#p332176

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 2:13 pm
by JohnChodera
That's weird! core22 0.0.5 was released to full FAH a week ago. You should be able to pick it up without problems.
Can you try deleting your core and work directories and seeing if that picks up the new core?
If not, we'll investigate ASAP!

~ John Chodera // MSKCC

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 2:23 pm
by HaloJones
cores are stored in %APPDATA%\FAHclient\cores

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 2:29 pm
by scm2000
JohnChodera wrote:That's weird! core22 0.0.5 was released to full FAH a week ago. You should be able to pick it up without problems.
Can you try deleting your core and work directories and seeing if that picks up the new core?
If not, we'll investigate ASAP!

~ John Chodera // MSKCC
I'll do that as soon as my other 2 work units complete.

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 2:42 pm
by astrorob
i have a machine with 3 GPUS, one fast, one middling and one slow.

the two faster GPUs are showing state "UPDATE_CORE" in fahclient. they are not requesting work units.

the slower GPU is still working on a WU which will take 2 more days to complete.

i think this means that the client is waiting for the FahCore22 that's still running to finish before trying to update the binary. does that sound right? i'm going to waste 2 days of 2 faster GPUs if this is true. is there a way for me to abandon the WU on the slow GPU and cause the core update to complete?

thanks!

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 2:55 pm
by astrorob
well, it was enough to pause the slow GPU. for some reason i then got an error in one of the slot logs saying the core binary could not be overwritten but maybe that's because another slot was writing it. the two slots waiting on UPDATE_CORE then immediately started up on 0.0.5.

the slow GPU checkpoint loaded OK on the new core binary after unpausing.

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 3:14 pm
by scm2000
it makes sense it would not update the core until the other work units complete...
ah but I did pause everthing, they all went to update_core.. then unpausing it updated, then they all started running again.

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 5:11 pm
by Joe_H
Yes, this is an issue with more than one folding slot using a core. All need to be inactive before a core update will proceed.

Re: Update_Core status on a work unit.

Posted: Mon May 04, 2020 7:10 pm
by JohnChodera
We've heard reports that you can pause the active WUs and resume them with the updated core just fine. Might be worth trying if you don't want to wait!

~ John Chodera // MSKCC