High Performance Computing
at Memorial University

This page is near obsolete. The AS4100 project has now been incorporated into the MUNARCC (Advanced Research Computing and Communcations) group and the Researc Computing Steering Committee. Please refer to HPC Resources web page.

Table of Contents

[Digital AlphaServer 4100]
[Access Policies] [Access]

Digital AlphaServer 4100

[Digital As part of Digital Equipment Corporation's commitment to HPCnet, Digital has selected Memorial University of Newfoundland to host an AlphaServer 4100 for use by the Memorial University community, members of HPCnet, and other users in Newfoundland and Canada. The AlphaServer 4100 is a symmetric multiprocessor computer with four 400 MHz Alpha processors and 4 GB of RAM. This system was originally on loan to Memorial through an agreement with Digital, but has since been purchased.

Monthly usage reports of the AlphaServer 4100 are available.

Selected Benchmarks: (source: Digital alpha performance).
SPECint_rate95
SPECfp_rate95
LINPACK 100x100
LINPACK 1000x1000
422
439
188.5 MFLOPS/CPU
1841 MFLOPS

Application Procedure and Conditions

Please refer to the HPC Resources Page.

Users will have login access to another Digital Alphasystem which will be used as a front end to access the 4100. From here, one can ftp files to/from your home site, compile code and submit batch jobs via LSF to the 4100.

Access Policies

Usage policies will be determined by a Steering Committee appointed by Memorial's Vice-President (Research), in consultation with the executive committee of HPCnet and users. The Steering Committee recognizes the following as guiding factors.

  1. Full utilization of the system is a priority.
  2. Priority should be given to jobs which make use of capabilities which are not otherwise available to individual users, e.g. the large memory and/or 4 processor architecture.
  3. Encouraging creative use of the unique capabilities should be a priority.
  4. Many users require access to the additional compute power, irrespective of the special architecture.
  5. Some jobs will require the full system for some periods of time.
  6. Setting of priorities and job scheduling will be done via LSF, and will be subject to the inherent degree of flexibility of LSF.
  7. System resources should be shared equitably.
  8. Users will be required to provide some characterization of their work and their requirements.

Accessing and Using the System