Hi there!
Thank you so much for contributing to the Fold for COVID program.
The error you are getting means the work queue ran out of jobs that run on devices with lower amounts of memory. One of your two devices managed to run some jobs, but there were no other jobs at the moment for your second device.
You can read in our FAQs this project supports devices with 1GB of RAM (or greater).
Please don’t give up, your device will sit idle until jobs that have lower memory requirements are submitted by researchers. It’s still great to have your device online and available for work!
@georgiats
Thanks for that input! The messages from the GUI would seem to mean that there is never going to be any work availble, so it is a bit confusing.
Perhaps, and occasional update message could be displayed to indicate that the rpi had polled the server but did not get any work? Otherwise, it just seems to be dead…
Keep up the good work - this is inspirational stuff ou guys are doing!
Most respectfully,
G
I now have 4 RPI 3B+'s set up and connected and all four have been idle for nearly 24 hours for lack of tasks accessible to a machine with only 1GB of RAM.
Perhaps there simply isn’t enough low spec computation needed in protein folding? After all, it is normally done on super-computers.
I’m wondering if there’s no way to cluster my RPi’s and make them more useful?
@PelicanMedia the BOINC client reads the available memory directly from /proc/meminfo; that reading is an indication of physical memory available on the device (minus GPU memory) rather than what’s currently free.
@chrisys@vinntec Thanks for the quick responses.
I have been using Raspberry Pis on and off since the first Pi 1 B back in 2012 and always bought from recommended distributors with no issues. The first time I buy one from Amazon, I get sent the incorrect one. This is indeed a 1GB and NOT a 4GB, as I ordered.
Sorry for the confusion and the explanation @chrisys.
I guess I’ll let my 8 Titan X’s do the heavy work on the F@H project and have this Pi wait for a lower job to chunt through
Oh wow that’s annoying @PelicanMedia! I guess it’s too late to return the 1GB one and get the 4GB one you paid for now? Thanks for trying the project and leaving the 1GB device online anyway
Well, my 4 rpi 3B+'s hae been sitting idle for a few days now. Is there any website where we can see the outstanding task queue and the requirements for each?
I’m probably going to take my machines off line since they are not helping…
Maybe you can try to cluster them the way you would do with normal Raspbian?
After that you could try this idea: https://pimylifeup.com/raspberry-pi-boinc/
Have no idea if BOINC for arm supports Rosetta@home but it is worth a shot :).
Hello all.
I have a Pi4 with 4gb ram. it has been running this software for several days. It seems to stay rather busy. I notice that the Storage free seems to be reducing a good bit, My guess is the finished tasks are not erased after they are uploaded.
BUT! I also note that boinc is using 4 to 5 gb. SO, if you are using a smaller micro sd card, try a Larger card. I am using a 32 gb card, and would suggest the same.
The example screen not only does not have enough ram for some tasks, it also does not have enough storage for some tasks.
If you get no tasks, try a bigger storage card.
@Cauliflower12432@redpcx welcome to the community and thanks for the tip about storage! That is useful to know and may be something we can look into to make sure old tasks are removed.
My Pi 4 4Gb works very well. While also my Pi 3B + is instead stopped in the absence of light tasks.
Is it not possible to use ZRam to “increase” the available RAM to accept the calculation tasks?
Hi @Zzed – we’ve done some experiments on that internally, but haven’t come to any conclusions about performance. If you’d like to give it a try, it would be great if you could report your results back to the forum.