clGetPlatformIDs() returned -1001

Moderators: Site Moderators, FAHC Science Team

tripplehelix
Posts: 8
Joined: Sun Dec 10, 2023 4:05 pm

Re: clGetPlatformIDs() returned -1001

Post by tripplehelix »

@bikeaddict

That’s fascinating, so nothing I've done has actually counted? I'll power down my Pi. I was under the impression that I was helping as I was receiving points. Very sad to see, unfortunately I don't have any other free hardware to run folding@home on. I'll repurpose this Pi.
bikeaddict
Posts: 210
Joined: Sun May 03, 2020 1:20 am

Re: clGetPlatformIDs() returned -1001

Post by bikeaddict »

This particular WU appears to have been counted. But it was returned 1.5 days later than the researcher wanted.

It's possible that the system did regenerate the WU to be reassigned to another client, but it wasn't sent out before you returned it. There could be a process that removes reassigned WUs that miss the timeout but come in before they expire. Or maybe the WU was actually sent out and run again, but that report doesn't show the second one being completed.

There is still a problem with the assignment process that needs to be addressed. Any hardware that is taking multiple days to finish a WU and frequently missing the timeout should only be getting assigned smaller 1-day timeout WUs.
tripplehelix
Posts: 8
Joined: Sun Dec 10, 2023 4:05 pm

Re: clGetPlatformIDs() returned -1001

Post by tripplehelix »

Surely they should base future WU's based on what the hardware has managed in the past. A quick speed test of the machine the first time it connects to folding@home would suffice.
Joe_H
Site Admin
Posts: 7922
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: clGetPlatformIDs() returned -1001

Post by Joe_H »

tripplehelix wrote: Tue Dec 12, 2023 3:29 pm Surely they should base future WU's based on what the hardware has managed in the past. A quick speed test of the machine the first time it connects to folding@home would suffice.
That would require keeping statistical data of tens of thousands different systems and an increase in complexity to the assignment process that a single developer is not going to be able to support. That "quick speed test" could then be invalidated by the next configuration change made by the user or by the user not folding full time.

They do take reports of hardware not being able to consistently complete WUs from various projects and do have assignment parameters to keep a project from being assigned to ARM for instance. Or larger GPU based projects may only get assigned to th more powerful GPUs.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
toTOW
Site Moderator
Posts: 6349
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:

Re: clGetPlatformIDs() returned -1001

Post by toTOW »

bikeaddict wrote: Tue Dec 12, 2023 12:41 pm It's possible that the system did regenerate the WU to be reassigned to another client, but it wasn't sent out before you returned it.
Yes very likely. This process depends on the pressure put on the FAH system and this particular project or hardware combination ...

For this particular trajectory :
- WU has been assigned to a client that failed it immediately
- WU has been set out to tripplehelix
- WU expired and has been marked for resend
- tripplehelix returned a valid result before the WU has been reassigned, so it has been marked as valid and it has been used to create next WU
- next WU has been sent and completed

In this case, the work sent after timeout has not been wasted, but if there are a lot of clients fighting for this kind of work, it would gave been wasted ...
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
Post Reply