General Info and Policies
User Forms
UofA HPC User Policy
Form
Star Cluster Tutorial
Job scheduling with the Moab Workload Manager and the TORQUE Resource Manager
The Moab Workload Manager is a scheduler. Moab works with the TORQUE
resource manager. TORQUE manages the job queue and manages the compute
nodes. Moab tells the resource manager what to do, when to run jobs,
and where. Users typically submit jobs and query the state of the
machine and jobs through the workload manager.
Queues on Star:
parshort
- 8 core minimum, 72 hour max run time, higher priority than parlong.
parlong - 8
core minimum, 21 day max run time.
serial - 1-2
core jobs, no max run time limit.
debug - 16
core max limit, 1 hour max run time.
Sample Job Scripts:
LaMMPS
PQS
OpenMP
Serial job with
1 processor
Short parallel
job (<72 hour run time)
Long parallel
job (<21 day run time)
Memory intensive
parallel job
Jobs that need access to large
local disks.
One node, eight core example
script
Two node, sixteen core example
script
Three node, twenty-four core example
script
Two node, ten core example
script
Four node, sixteen core example
script
Moab and TORQUE
Documentation
Moab
User's Manual
TORQUE
User's Manual
Compiler Documentation
Gnu 4.1.1 Compiler Docs
Gnu C Users Manual
Gnu C++ Users Manual
Gnu Fortran
Users Manual
Intel 10.1 Compiler Docs
Optimization Quick
Reference
C
/ C++ Users Manual
Fortran
Users Manual
MPI programming guide and References
The Message Passing Interface Standard (MPI) is a message passing library standard based on the consensus of the MPI Forum, which has over 40 participating organizations, including vendors, researchers, software library developers, and users. The goal of the Message Passing Interface is to establish a portable, efficient, and flexible standard for message passing that will be widely used for writing message passing programs.
MPI Resources
Open MPI Website
MPI
Tutorial
MPI
Routines
OpenMP programming guide and References
OpenMP is an Application Program Interface (API), jointly defined by a group of major computer hardware and software vendors. OpenMP provides a portable, scalable model for developers of shared memory parallel applications. The API supports C/C++ and Fortran on multiple architectures, including UNIX & Windows NT.
OpenMP Resources
OpenMP
Overview
OpenMP
Tutorial
OpenMP Specifications
Last updated: August 28, 2009
