Open cstackpole opened 8 years ago
8/22/2016 8:09 AM, cstackpole пишет:
Greetings,
Looking for more information please.
Hello, cstackpole!
I have a SLURM cluster that has a bunch of downtime for a while as a project ramps up. I would like to put the nodes to use through the scheduler.
Are you still using this?
Yes, currently it is launched of one cluster.
What version of SLURM are you currently using?
It is rather old, I can check it later.
What version of BOINC clients?
7.0.65 In fact you can use any version of BOINC.
Can this work with other projects, or is it just SAT?
Yes, sure.
Can this project utilize multiple partitions (different types of hardware / cpu count / memory)? What about GPU?
There are no such limitations. CluBORun just helps to launch BOINC client on several cluster nodes. All other things are performed by BOINC client itself. You can run install.sh from here https://github.com/Nauchnik/CluBORun/tree/master/SLURM_install After it you should compile and link BOINC on your cluster. I will tell you about the next steps. By the way, I will do my best to make a manual in english
Best regards, Oleg Zaikin.
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
Awesome! Thank you very much. I am close to having the time to test it out and I am looking forward to trying. Thanks!
Greetings,
I ran the document through an online translator. It kinda worked. It looks like the repo is an older version of 32bit Boinc I am trying to re-compile a Boinc version that is 64bit on Scientific Linux 6. That is proving to be a bit more challenging. :-)
I am curious what cluster OS/arch you are running this on. Have you updated the Boinc version since?
Thanks!
Hello! We did not upgrade BOINC because it is not required for us - we use our cluster only for CPU computing.But this version of BOINC we use due limitations of OS and set of libraries on this cluster. You can non use it.If usual version of BOINC runs on your cluster without any problems, you can simply use it instead of "our BOINC" which we compiled for our cluster. CluBORun is a simple set of scripts with main functionality in catch_node.awk for analyzing status of task queue and set (or break) BOINC-tasks into it. It "catch" node (or several nodes) by task and starts BOINC client on they. You absolutely free in version of BOINC that you can use. 17.09.2016, 21:23, "cstackpole" notifications@github.com:Greetings,I ran the document through an online translator. It kinda worked. It looks like the repo is an older version of 32bit Boinc I am trying to re-compile a Boinc version that is 64bit on Scientific Linux 6. That is proving to be a bit more challenging. :-)I am curious what cluster OS/arch you are running this on. Have you updated the Boinc version since?Thanks!—You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or mute the thread.
Greetings,
Looking for more information please.
I have a SLURM cluster that has a bunch of downtime for a while as a project ramps up. I would like to put the nodes to use through the scheduler.
Are you still using this?
What version of SLURM are you currently using?
What version of BOINC clients?
Can this work with other projects, or is it just SAT?
Can this project utilize multiple partitions (different types of hardware / cpu count / memory)? What about GPU?
Thanks!