Texas A&M Supercomputing Facility Texas A&M University Texas A&M Supercomputing Facility

General Comments, Caveats, & Recommendations

  1. Since about mid-April 2007, we have started doing daily backups of all active user directories in the /home area only. The rest (e.g., /scratch/username/*) is your responsibility to move to our tape archive or elsewhere, if need be.
  2. Know your program. Know, approximately that is, how much memory, how much I/O and disk space, how much CPU and wall-clock time it needs for a run. If it runs in parallel mode, it is important to know how well your program's performance scales as you increase the number of processors. At the very least try to find out. We will help. Also, be aware that in submitting parallel MPI batch jobs there are several ways to indicate how you want your job distributed across the available nodes. Read more on this important topic in the subsection on batch jobs, "Task Distribution across Nodes".
  3. Prefer 64-bit executables. Always prefer to construct 64-bit executables, if you have a choice. These, and not the 32-bit versions, include all the high performance features of the Power5+ architecture, as well as allow for a very large memory address limit. With 32-bit executables you're stuck at 4 gigabytes. But this limit is frequently exceeded when solving large problems. It is for these reasons that we consistently use the -q64 compiling option below. By the way, this 64-bit issue has nothing to do with numeric data precision.
  4. Prefer using only Bourne shell compatibles. For your scripts, please use only Bourne Shell compatibles or derivatives: ksh (Korn), sh (Bourne). The Korn shell is the standard Unix shell on hydra. Also, even though generic bash is in many ways desirable, on AIX it has some pesky problems.
  5. We strongly discourage you from transfering startup scripts (.login, .profile, etc.) from other systems, or casually changing the ones that your account comes with, unless you understand the effects of those changes. And don't expect PBS or LSF batch scripts to work on this cluster. This admonition also applies to configuration files related to various applications, such as the ones used by ABAQUS, CFX, etc. If you need help let us know.