Page 1 of 1

Project: 18011 has too short deadline

Posted: Fri Jul 16, 2021 5:53 am
by vmzy
I had jsut received a Project: 18011 (Run 0, Clone 404, Gen 19) which has 1 day Timeout and 2 days Expiration.
But on my laptop(with mx150) its TPF is 19.5 minutes,ETA is 1.37 days. which means that I can`t finish it befor Timeout even if I ran it 7*24.If i shutdown my laptop overnight the WU will be expired.
PLZ extend the deadline of this Project TKS.

Re: Project: 18011 has too short deadline

Posted: Fri Jul 16, 2021 8:14 am
by vmzy
I noticed rafal.wiewiora have set all its GPU Project to 1 day Timeout and 2 days Expiration.Even Project 18010 with 209,848 base points. my Project 18011 has 96,016 base points.

Re: Project: 18011 has too short deadline

Posted: Fri Jul 16, 2021 9:05 am
by aetch
It sounds like the project needs tweaked to raise the minimum GPU species it's allowed to run on.

I doubt if you'll receive the higher value work units, there's quite a few desktop cards that would struggle to process them and the minimum GPUs species should already be very high.
I think part of the problem is that when the GPUs are assigned a species it's based upon capability, not performance. The performance difference between a desktop and mobile variant of the card can be considerable.
Those higher value units can be processed well within the 24 hour timeout but it takes high-end kit. My own RTX 2070 Super recently munched through a 113K unit in about 4 hours.

I don't run the lower power stuff anymore. There was a time I didn't mind adding it in part-time to give a boost with smaller units but it kept being hit with big units, which really bugged me because my high-end kit was only seeing the smaller units. I don't mind getting the big units but would have preferred to see it on the high-end kit.

Re: Project: 18011 has too short deadline

Posted: Fri Jul 16, 2021 3:26 pm
by Joe_H
Or possibly the MX150 needs to be reclassified to a lower species level that it currently is. That and a few other of the low end GPUs.

A project was started last year to change how WUs from various projects were assigned to GPUs of different processing speeds. It was paused for a while when they needed the developer resources elsewhere, so I don't know when they will complete this change.