User Tools

Site Tools


techstaff:introduction

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision
Previous revision
Next revisionBoth sides next revision
techstaff:phd_orientation_2017 [2017/09/19 11:09] – [Self Management] kauffmantechstaff:phd_orientation_2017 [2017/09/19 11:13] – [SLURM] kauffman
Line 28: Line 28:
 =====Services===== =====Services=====
 ====Interactive login servers==== ====Interactive login servers====
-Techstaff has 3 servers (linux1, linux2, and linux3) that serve as interactive login systems for instructional use.  These systems are reachable either by the canonical address or by a cname (linux.cs.uchicago.edu).  The address linux.cs.uchicago.edu is a DNS round-robin entry that will rotate between the canonical addresses.  This allows for future expansion of the service.  You should always be aware of the canonical in case you need to reconnect to a specific machine.+[[techstaff:remote_access|Techstaff has 3 servers (linux1, linux2, and linux3)]] that serve as interactive login systems for instructional use.  These systems are reachable either by the canonical address or by a cname (linux.cs.uchicago.edu).  The address linux.cs.uchicago.edu is a DNS round-robin entry that will rotate between the canonical addresses.  This allows for future expansion of the service.  You should always be aware of the canonical in case you need to reconnect to a specific machine.
 These systems are configured similarly to the desktops and lab machines and can be used interchangeably with them.  They are configured with 16 hyperthreads, 64GB of RAM, 450GB /local partition, and mount the NFS home directory. These systems are configured similarly to the desktops and lab machines and can be used interchangeably with them.  They are configured with 16 hyperthreads, 64GB of RAM, 450GB /local partition, and mount the NFS home directory.
 The purpose of these systems is to support instructional computing.   The purpose of these systems is to support instructional computing.  
Line 68: Line 68:
  
 =====SLURM===== =====SLURM=====
-The SLURM cluster was originally constructed to support Hank Hoffmann’s OS course.  SLURM is a job scheduling system that allows users to submit requests for resources and allows them to execute jobs using those reserved resources.  (Information about SLURM and its use is available at https://howto.cs.uchicago.edu/techstaff:slurm.)   +The [[techstaff:slurm|SLURM cluster]] was originally constructed to support Hank Hoffmann’s OS course.  SLURM is a job scheduling system that allows users to submit requests for resources and allows them to execute jobs using those reserved resources.   
-The cluster can also be rather than linux.cs.uchicago.edu for larger jobs Additionally, we have a couple of GPUs that can be utilized on the cluster.+The cluster should be used for larger jobs, rather than than linux.cs.uchicago.edu. Additionally, we have a couple of GPUs that can be utilized on the cluster.
  
 =====Hadoop===== =====Hadoop=====
Line 100: Line 100:
 Create a directory within /local on the system that you are using for data files that are large or will be accessed regularly by a program.  Doing so will improve performance (local disk is much faster than network disk) and will bypass home directory quotas. Create a directory within /local on the system that you are using for data files that are large or will be accessed regularly by a program.  Doing so will improve performance (local disk is much faster than network disk) and will bypass home directory quotas.
  
-To do so: 
 <code> <code>
 cd /local cd /local
/var/lib/dokuwiki/data/pages/techstaff/introduction.txt · Last modified: 2023/09/07 14:28 by adahler

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki