Discussion in 'Team OcUK Distributed Computing Projects' started by Biffa, Nov 24, 2018.
wow only 8 years till we catch tech report, cant wait till we stomp them
Who is Add77 on the forum does anyone know? Can't find them listed here, must be using a different handle. Seems they have similar setup to me.
not sure tbh. I can believe a billion points doesnt even get you in the top 400. it used to be top 100 not so long ago.
Whats your folding setup Biffa?
Just a couple of 2080Ti's
Cool. Theoretical situation here. Assuming everything remaining equal, how many 2080ti’s would it take to get us in the top 10 within a year!?
More than 60
Any way of getting a notification when one of your machines stops folding?
One of my VMs has been stuck downloading a WU since 26/11 which has wasted 100K points a day.
Incidentally it's nice that they've restored some of the credit for CPU folding, as I have no GPUs I can fold on.
Hmm, is it a Linux VM?
I only ask because I had to write a windows batch file to restart my folding on one of my computers because twice a day it got stuck downloading. So I wrote a batch file to monitor the time stamp of the log and if it didn't change in 3 minutes to kill and restart all the fah processes.
I've got some Windows desktops and a couple of those have got stuck downloading, but it's the 2 Linux VM's that are the big hitters, they average between 100K and 300K PPD each.
I've got FAHControl setup to remotely monitor all of them, but it's not ideal because they all just show 'Online' as a status rather than the actual status of what the Client is currently doing (Running, Downloading, Paused), so I have to click through the whole list and wait for it to poll each one to find out if they're stuck.
The scripts a great idea, I'll knock one up in shell, thanks.
With cpu folding you might want to adjust the watch time a bit longer, maybe 5 minutes.
Yeah, I'd noticed the timestamps in the log file were slightly further apart than for GPU, I'll investigate further when I eventually get time.
Are there any flags I should be running to increase PPD or have they all been discontinued?
for cpu or gpu? not sure about cpu as I only run gpu
CPU, no worries.
I had a read through the docs but it seems to imply that the flags have been discontinued.
Yeah, I just tried one in Virtualbox on my Threadripper got about 180K using 30 threads, just did the client-type advanced
The client-type advanced flag doesn't get you any extra PPD according to the docs, It can reduce it if you get a less stable WU.
It used to be bigadv, but it looks like this ended a long time ago.
Due to the above I have just left my clients at the default settings.
Thanks for doing the news Biffa, sorry for my lack of participation recently, work and life taking up time
Separate names with a comma.