31st of July

Marks Rpi Cluster is running 24/7. After a break to migrate the Pi4's from their cases to the Edge Cluster 12 they are now running flat-out doing Einstein BRP4 work. Einstein had the BRP4 work turned off at the beginning of the week so I ended up running their FGRP5 work which were taking 25 hours a work unit. The BRP4's are much more digestible at 3 hours and 5 minutes.

BitScope do make larger versions than the EC12, but that was all the Pi4's I had on hand and its a reasonable size to deal with. Most of the Pi's are running between 53 and 59 degrees C so airflow doesn't seem to be an issue. That brings Mark Rpi Cluster to the following configuration:

12 x Pi4 8GB as compute nodes (in the EC12).
3 x Pi4 2GB as support nodes.

I have ordered another couple of Meanwell 24v power supplies. I am thinking I might get another EC12, assuming I can find some Pi4's to put in it. The Pi4's are out of stock everywhere at the moment.