1 / 10

Experiences of Grid Enabled MPI Implementation named MPICH-GX with Atmospheric Applications

Experiences of Grid Enabled MPI Implementation named MPICH-GX with Atmospheric Applications. Oh-kyoung Kwon, KISTI Salvador Castañeda, CICESE. PRAGMA 11. Contents. Motivation KISTI’s Proposal MPICH-GX Experiments Next things to do. Atmospheric Applications. MM5

oleg
Download Presentation

Experiences of Grid Enabled MPI Implementation named MPICH-GX with Atmospheric Applications

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Experiences of Grid Enabled MPI Implementation named MPICH-GX withAtmospheric Applications Oh-kyoung Kwon, KISTI Salvador Castañeda, CICESE PRAGMA 11

  2. Contents • Motivation • KISTI’s Proposal • MPICH-GX • Experiments • Next things to do PRAGMA 11

  3. Atmospheric Applications • MM5 • A weather model designed to simulate or predict atmospheric circulation. • WRF • A next-generation numerical weather prediction system designed to serve both operational forecasting and atmospheric research needs. PRAGMA 11

  4. The Problems • Global simulations require a great amount of resources and processing power • Gigabytes of hard disk space for distributed databases • Gigabytes of RAM • Terascale computation PRAGMA 11

  5. The Problems (cont'd) • Running on a Grid presents the following problems: • Standard MPI implementations require that all compute nodes are visible to each other, making it necessary to have them on public IP addresses • Public IP addresses for assigning to all compute nodes aren't always readily available • There are security issues when exposing all compute nodes with public IP addresses PRAGMA 11

  6. KISTI's Proposal • To solve the problem between public and private IP addresses, KISTI have developed MPICH-GX PRAGMA 11

  7. MPICH-GX • MPICH-GX is a patch of MPICH-G2 to extend functionality required in the Grid. • MPICH-GX provides following functions. • Private IP Support • Fault Tolerant Support • Currently, MPICH-GX can run with GT3.x  In this year, we might move to GT4. PRAGMA 11

  8. The Testbed eolo pluto * The job on each cluster is scheduled through OpenPBS. PRAGMA 11

  9. Experiments • 12 nodes on single cluster (orion cluster): 17.55min • 12 nodes on cross-site • 6 nodes on orion + 6 nodes on eros, where all nodes have public IP: 21min • 6 nodes on orion + 6 nodes on eros, where the nodes on orion have privateIP and the nodes on eros have public IP: 24min • 16 nodes on cross-site • 8 nodes on orion + 8 nodes on eros, where all nodes have public IP: 18min • 8 nodes on orion + 8 nodes on eros, where the nodes on orion have privateIP and the nodes on eros have public IP: 20min PRAGMA 11

  10. Next things to do • Test with larger input data • Study on minimum requirements of network bandwidth for MPI jobs in Grid environments • Working with MPICH-G4 and GT4 GRAM • Extension to include PRAGMA testbed PRAGMA 11

More Related