Ability to abort/cancel a workunit?

Moderators: Site Moderators, FAHC Science Team

Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

Yes that makes sense, it's been set up to get it done as quick as possible.

So if I happen to go over the time limit, I should leave it running. Will my FAH client automatically cancel it when it reaches the expiry?

What times don't make sense to you?
Neil-B
Posts: 2027
Joined: Sun Mar 22, 2020 5:52 pm
Hardware configuration: 1: 2x Xeon E5-2697v3@2.60GHz, 512GB DDR4 LRDIMM, SSD Raid, Win10 Ent 20H2, Quadro K420 1GB, FAH 7.6.21
2: Xeon E3-1505Mv5@2.80GHz, 32GB DDR4, NVME, Win10 Pro 20H2, Quadro M1000M 2GB, FAH 7.6.21 (actually have two of these)
3: i7-960@3.20GHz, 12GB DDR3, SSD, Win10 Pro 20H2, GTX 750Ti 2GB, GTX 1080Ti 11GB, FAH 7.6.21
Location: UK

Re: Ability to abort/cancel a workunit?

Post by Neil-B »

Yes the client will drop a wu if it passes expiry deadline ... timeouts and expiry deadlines are set by the researchers based on various factors relating to urgency of research, types of project, and sometimes they can be a bit obscure ... a small fast wu with long deadlines - a large long wu with short deadlines - a project with a shorter number of gens may be less time sensitive than ones with lots of gens or sometimes not ... basically the researchers set the deadlines and sometimes they can just seem odd to folders
2x Xeon E5-2697v3, 512GB DDR4 LRDIMM, SSD Raid, W10-Ent, Quadro K420
Xeon E3-1505Mv5, 32GB DDR4, NVME, W10-Pro, Quadro M1000M
i7-960, 12GB DDR3, SSD, W10-Pro, GTX1080Ti
i9-10850K, 64GB DDR4, NVME, W11-Pro, RTX3070

(Green/Bold = Active)
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

All those make sense to me, it depends if they need things quickly, or they want to utilise resources without wastage.

P.S. what are you doing with that 512GB RAM?
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

Joe_H wrote:Basically this is a borderline CPU, enough to handle the CPU side of GPU processing but will not complete all CPU WUs within deadlines. Without AVX the CPU folding core uses SSE2 instructions and runs somewhat slower than on a system with AVX available. On a faster processor this is good enough, I ran folding on a 12 year old processor that only supported SSE until its PS failed, and never ran into problems completing WUs within their deadlines.
Pah! Not everybody has super fast processors. That machine allowed me to play games! Using the onboard graphics! Fine, I'll use the Ryzen 9 3900XT....
gchristopher
Posts: 11
Joined: Thu Apr 30, 2020 2:45 pm

Re: Ability to abort/cancel a workunit?

Post by gchristopher »

It's a shame I can't cancel. I'll give you my use case.

This machine is usually booted into a different OS (unusual) but I had some time to do a long disk copy (10 hours) so I decided to allow it to fold.

Both a 24 CPU and a 32 CPU core (my machine has 56 virtual cores) Slot each got a 3 day job. Imagine how big these are for 24 cores to take 3 days.

The problem is that I'll be done copying in another 8 hours, then booting back to other system.

Since both jobs will take 3 days to complete and there is no way I can do them, I cancelled them.

Problems:
1) I can't fish for a shorter block. So 56 cores will go unused.
2) The researcher gets burned by having 2 very large jobs delay quite a while until they timeout.

Bummer for everyone.
calxalot
Site Moderator
Posts: 878
Joined: Sat Dec 08, 2007 1:33 am
Location: San Francisco, CA
Contact:

Re: Ability to abort/cancel a workunit?

Post by calxalot »

I believe you can delete their slots in FAHControl and the WUs will get dumped properly.
It will still count as failures against your quick return bonus success rate.
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

gchristopher wrote: Wed Oct 18, 2023 2:10 amBoth a 24 CPU and a 32 CPU core (my machine has 56 virtual cores)
What do you mean? How many are real cores? No point in putting two programs onto one real core.
calxalot wrote: Wed Oct 18, 2023 5:55 amI believe you can delete their slots in FAHControl and the WUs will get dumped properly.
It will still count as failures against your quick return bonus success rate.
Just pause the computer in Folding, then dump the workunit you don't want with the bin button. You may lose bonus points if you do it too often, but you have to screw up 20% of them. If you can't do it, best to let the system know earlier so it can go to someone else and not delay the research.
calxalot
Site Moderator
Posts: 878
Joined: Sat Dec 08, 2007 1:33 am
Location: San Francisco, CA
Contact:

Re: Ability to abort/cancel a workunit?

Post by calxalot »

The ‘bin’ button is in v8.
I believe the recent poster is using v7.
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

Oh, I assumed both versions had the bin button. I also assumed we were all on the latest version. 8 is so much better than 7.
Joe_H
Site Admin
Posts: 7856
Joined: Tue Apr 21, 2009 4:41 pm
Hardware configuration: Mac Pro 2.8 quad 12 GB smp4
MacBook Pro 2.9 i7 8 GB smp2
Location: W. MA

Re: Ability to abort/cancel a workunit?

Post by Joe_H »

No, v7 doesn't have the bin button. Dumping a WU is possible several ways, none directly like that. v8 is still a public beta and has some missing functionality and other issues to be resolved before becoming a full release version. Development continues behind the scenes.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

I've found nothing missing from 8. What is it I can't do? All I've found is it's much easier to use.
Joe_H
Site Admin
Posts: 7856
Joined: Tue Apr 21, 2009 4:41 pm
Hardware configuration: Mac Pro 2.8 quad 12 GB smp4
MacBook Pro 2.9 i7 8 GB smp2
Location: W. MA

Re: Ability to abort/cancel a workunit?

Post by Joe_H »

Well, for starters v7 by default will pause WU processing if a laptop loses AC power and is running on its battery. Depending on how connected and the OS, the same logic will also pause a desktop connected to a UPS. That is not available at all in v8.

Without an active internet connection the v8 web control does not work. Then you can not stop, start or pause processing by the client. There are some workarounds, but no easy way to control it yet under this circumstance.

There are other problems, I would have to go back and find the list I put together a couple months ago.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

I'd also add having to bypass security on the browser may be something many folk are not prepared to do. I don't really understand it, Chrome tells me the site is insecure, even though I've told it to allow insecure parts. Yet if I click on it, it says the certificate is valid, so what's the problem? If I change it to the default of blocking insecure parts, it still tells me it's insecure - how can it be if those parts are blocked?
calxalot
Site Moderator
Posts: 878
Joined: Sat Dec 08, 2007 1:33 am
Location: San Francisco, CA
Contact:

Re: Ability to abort/cancel a workunit?

Post by calxalot »

It’s because a secure page (https) is opening an unencrypted websocket to localhost (to the local client).
Peter_Hucker
Posts: 306
Joined: Wed Feb 16, 2022 1:18 am

Re: Ability to abort/cancel a workunit?

Post by Peter_Hucker »

Chrome's messages are nonsensical, if I've blocked the unsecure connections, why does it have anything to say? If I haven't blocked them, why isn't it telling me what you just have?

So what's the way round this? The clients have to do https too? I wish https had never been invented, it causes no end of problems over at Boinc too, when somehow it's ok for a whole year but then 2 seconds out of date is dangerous. Kinda pedantic like the car MOT (annual safety test).
Post Reply