National Partnership for Advanced Computational Infrastructure: Archives

These pages are a copy of the original www.npaci.edu website, and should be used for historical reference only.
Please select an item from the toolbar below to be taken to the latest information on that subject.
[ SDSC | User Services | Applications | Allocations | Consulting | SAC | Datastar | Training ]


NPACI Grid: Resources


ABOUT NPACI Grid
What Is It?
Case Studies
Grid Monitor
Testbed Info
Terminology
FAQ

USER REFERENCE
Getting Started
Tutorial
Certificates
Resources
NPACKage
HotPage

LEARN MORE
Events
Web Links
Contacts

 

NPACI Archive Page

The NPACI program ended on September 30, 2004. This site is presented for archival purposes only. For current resources at each of the partner sites, please refer to the appropriate institution site.

User Guide - Resources

This document describes the resources at SDSC, Texas, and Michigan that make up the NPACI Grid.  Following this summary, a table of NPACI Grid login hosts is provided, followed by a a Grid Services Matrix specifying the services available on the NPACI Grid.

If you have an NPACI account but do not have access to one or more of the NPACI Grid resources, submit an account extension request at:

http://npacigrid.npaci.edu/account_extension_request.html

SDSC

Blue Horizon

Blue Horizon is a teraflop-scale Power3 based clustered SMP system from IBM.  The machine contains 1,152 processors and 576 GBytes of main memory, arranged as 144 Symmetric Multiprocessing (SMP) compute nodes.  Each node is equipped with 4 GBytes of memory shared among its 8 - 375 MHz Power3 processors.  Each node also has several GBytes of local disk space.  Nodes are connected by the Colony switch, a proprietary IBM interconnect.

Although most of the Blue Horizon resources are batch systems, 15 nodes are available for interactive use.  These "b80 nodes" are for developing and debugging parallel codes and are not intended for making production runs.  For more information:

Blue Horizon Home Page
System Access
Running Batch Jobs
Running Interactive Jobs

Griddle

Griddle is the host where Condor-G jobs must be submitted on the NPACI Grid. From griddle. jobs may be launched on any NPACI Grid resource.  Griddle has an Intel Pentium 4 IA32 processor and runs Redhat Linux version 7.2. 

Texas

Longhorn

The TACC IBM Power4 System consists of three IBM p690-HPC shared-memory server nodes, one IBM p690-Turbo shared memory node, and 32 IBM p655 shared-memory server nodes along with an IBM p690-HPC node as a login front-end.  Each of the p690-HPC server nodes contains 16 Power4 processors running at 1.3 GHz, while the p690-Turbo contains 32 Power4 processors at the same speed.  Each of the p655 server node contains 4 Power4 processors running at 1.3 GHz.  In total, the 224 total processor system has a peak performance 1.16 Tflops and an aggregate memory of 512 GB.  Each node is supported by 36/18 GB of local disks with an aggregate of 3/4 TB, and the faster storage GPFS is connected through the IBM Switch2 for a total of 7.1 TB of storage disk space.  The Power4 Systems run AIX, a scalable UNIX operating system with High Availability Cluster Multi-Processing (HACMP) capabilities.  For more information:

Longhorn Home Page
System Access
Running Batch Jobs
Running Interactive Jobs

Michigan

Hypnos and Morpheus

The hypnos cluster has 128 nodes of dual Athlon 2000MP CPUs and is reserved for NPACI users.  The morpheus cluster is for general use and consists of 50 nodes of dual Athlon 1600MP CPUs plus 17 nodes of dual Athlon 2600MP CPUs.  Each SMP node consists of two CPUs with one gigabyte of memory available per processor.  These AMD clusters run Red Hat Linux 7.2 and 7.3 with the typical GNU and Linux tools and utilities installed.  For more information:

AMD Clusters Home Page
System Access
Running batch and interactive jobs

Login Hosts

The following table specifies the hosts where you can login and access NPACI Grid services

NPACI Grid Login Nodes
Resource Site Host Notes
Blue Horizon

tf004i.sdsc.edu
tf005i.sdsc.edu

Batch job submission only
b80n01.sdsc.edu to
b80n13.sdsc.edu
Interactive job submission only
Griddle griddle.sdsc.edu Condor-G job submission and monitoring
Longhorn longhorn.tacc.utexas.edu Batch & interactive job submission
archive.tacc.utexas.edu Long term storage
Michigan morpheus.engin.umich.edu
hypnos.engin.umich.edu
Batch & interactive job submission

 

Grid Services Matrix

This table below specifies the services and parameters necessary for using the NPACI Grid A key is provided at the end of the table.

NPACI Grid Services & Parameters

General (not site specific)

GIIS Server giis.npaci.edu
Port 2135
NWS Name Server Host nws.npaci.edu
Name Server Port 8090 (default)
Memory Server Host nws.npaci.edu
Memory Server Port 8070 (default)
Condor-G Job Submission & Monitoring Host griddle.sdsc.edu
Globus Job Submission Gatekeepers

tf004i.sdsc.edu and
tf005i.sdsc.edu for batch and interactive jobs

b80n01.sdsc.edu through b80n13.sdsc.edu for interactive jobs

Jobmanagers jobmanager-fork (default)
jobmanager-loadleveler (batch)
Required RSL params for loadleveler on the b80 nodes (queue=interactive)
(max_wall_time=45)
(environment=(MP_EUIDEVICE en0))
Required RSL params for loadleveler on the tf004i and tf005i nodes

(queue=normal)
(max_wall_time=45)

 

GridFTP Server tf004i.sdsc.edu
GSI SSH Server tf004i.sdsc.edu
Port 1022
Globus Job Submission Gatekeeper longhorn.tacc.utexas.edu
Jobmanagers jobmanager-fork (default)
jobmanager-loadleveler (batch)
Required RSL params for loadleveler

(queue=normal)
(max_wall_time=45)
(max_memory=10)

max_wall_time is in minutes; max_memory is in megabytes; set

GridFTP Server longhorn.tacc.utexas.edu
GSI SSH Server longhorn.tacc.utexas.edu
Port 1022
Globus Job Submission Gatekeeper hypnos.engin.umich.edu
Jobmanagers jobmanager-fork (default)
jobmanager-pbs (batch)
Required RSL params for pbs

(queue=route)
(max_wall_time=45)
(email_address=your@email)

max_wall_time is in minutes; the above value of 45 is an example

GridFTP Server hypnos.engin.umich.edu
GSI SSH Server hypnos.engin.umich.edu
Port 1022
GRAM Job Submission Gatekeeper morpheus.engin.umich.edu
Jobmanagers jobmanager-fork (default)
jobmanager-pbs (batch)
Required RSL params for pbs (queue=npaci)
(max_wall_time=45)
(email_address=your@email)
GridFTP Server morpheus.engin.umich.edu
GSI SSH Server morpheus.engin.umich.edu
Port 1022

Key

  • GRAM Job Submission: Services and parameters for submitting Globus jobs
    • Gatekeeper: The node where the globus GRAM server is running
    • Jobmanagers: Specifies the default job manager (usually fork) and any other jobmanagers available, such as those for submitting batch jobs
    • Required RSL Parameters: Specific parameters required for submitting jobs via the given job manager.  Job submission will fail if these parameters are not set.
  • GridFTP: Host where the GridFTP server is running, specified for grid ftp transfer commands
  • GSH SSH: Host where the GSI SSH server is running.  Note that this server uses a different port than SSH, which must be explicitly specified when running GSI SSH commands.