CARC web archive · Archived copy of ahpcc.unm.edu/Systems/HPC_Systems.html as captured by the Internet Archive on 15 August 2000. Links to pages that were not saved here go to the Wayback Machine. Wayback copy · All pages in this snapshot · Web archive index


News
Accounts & Allocations
HPC Systems Home
Help Desk
Documentation
Training
Supercluster Research




High Performance Computing (HPC) Systems at the AHPCC

The HPC Systems Group at the AHPCC is responsible for the design, installation, configuration and maintenance of hardware systems located at the AHPCC. Please follow the links below to find the home pages for each system.

 

AHPCC & Alliance:

Name
Procs.
Nodes
MHz
Processor
Vendor
Type

O/S

Interconnect
LosLobos
512
256
733
Intel Pentium III
Netfinity 5500
Linux
Myrinet
RoadRunner
128
64
450
Intel Pentium II
AltaCluster
Linux
Myrinet
Access Grid
Comp. Grids

AHPCC Systems:

Name

Proc.

Nodes

MHz

Processor Type

Storage

Operating System

Blackbear
32   550 Pentium III 144 GB Dual SMP Myrinet Linux cluster
Blackbear is for general use as a development and production parallel machine. It is a high bandwidth, low latency interconnect machine. It is very good with jobs that require a lot of message passing between the processors. Blackbear can be used for classes. "Embarrassingly parallel" appelettes that do not require a HBLL interconnect should use Cochiti, or utilize the Condor Job Submission Capability to submit jobs on the Turing Cluster or RCDE boxes.
Turing Cluster
24   600 Pentium III SMP Linux
The Turing Cluster is a group of 12 machines that serve dual purposes. The cluster acts as workstations for workshop attendees and also as a computation engine for naturally parallel jobs. These jobs are scheduled through Condor, which migrates the jobs to other machines when the cluster is used for a workshop.
Azul
32   550 Pentium III 360 GB Quad SMP Gigabit Ethernet Linux cluster
Azul is the hardware side of a research project between AHPCC and IBM Research. A cluster of 32 processors (8 4-way NetFinity nodes) is joined together with a cluster of 32 processor Winterhawk machines with a Scalable Graphics Engine (SGE) to create a parallel computing, rendering and displaying environment. Access is allowed only to users who have been approved for the project. Please contact Dr. Bob Ballance for more information.
Nambe
8   195 R10000 2 GB SGI Onyx/2 IRIX
Cochiti
4   100 R4400 SGI Onyx
Grants
1   166 PowerPC 604 AIX
RCDE
35   333 Pentium II Windows NT & Linux
Belen
2   195 MIPS R10000 SGI Octane
Quemado
1   166 PowerPC 604 AIX
Isleta
4   SGI Onyx

 

Web Editor
Last Modified: August 14, 2000 12:09