well, they are actually at 500 TeraFLOPS peak
and it only costs $59 million
Hello All,
We have upgraded marge, homer, and several of the workstations. Everyone should have ssh access again. I’m about issue a command that will force all users to change their password the next time they log in. Please do not use a password that you use for some other service. For example, IM passwords are often stored in insecure places so don’t use your IM password for your lab password!
Ron
I am shutting down services in response to http://it.slashdot.org/it/08/02/10/2011257.shtml text me or IM if you need to get something done and cannot get access.
Ron
Andy’s Jura boards appear to be a success! I’ve ordered more parts and cables to finish assembling this run. The Mouser order arrives 2/12, USB cables on 2/12, and Digi-Key 2/13. Just in time for Valentine’s Day!
Ron
Hello all,
I flashed another 32 nodes (n00-n07, n08-n15, n24-n31, and n56-n63) so all of these nodes have the very latest boot controller (console running at 115,200). I ran a simple pi calculation and it seems to be working. However, a few nodes are causing trouble (refuse to boot intially). Other than this problem, woohoo, half the cluster is operational.
Ron
Peter is back up – ready to be used by anyone, especially master’s students. It was crashing because of 2 bad disks.
Keep in mind that that some of the lab workstations (peter, lois, and marge) are very old and prone to failure. Any data, programs, and libraries not stored in your home directory or in /build is not backed up and will probably be lost someday.
Will
Just a head’s up… n55 has been upgraded. It currently is the node with an extra serial cable connected to stewie in the lab.
The bootctld is backwards compatible; however the default baud rate of the console is now 115,200.
Ron