PDSI at NERSC

Introduction

Welcome to the Petascale Data Storage Institute (PDSI) website at the National Energy Research Scientific Computing (NERSC) center. PDSI is funded by the U.S. Department of Energy's SciDAC (Scientific Discovery through Advanced Computing) program, and is a collaboration between researchers at Carnegie Mellon University, National Energy Research Scientific Computing Center, Pacific Northwest National Laboratory, Oak Ridge National Laboratory, Sandia National Laboratory, Los Alamos National Laboratory, University of Michigan, and the University of California at Santa Cruz. For more information on the collaboration, please go to the project website.
While there are numerous goals of PDSI, NERSC is currently focused on providing:
  1. Detailed failure data of various system and component problems.
  2. Information on I/O benchmarking and characterization.

Failure Data

Our goal is to provide failure specifics for I/O related systems and components in as much detail as possible so that analysis might produce some useful findings. Data were collected for storage, networking, computational machines, and file systems in production use at NERSC from the 2001-2006 timeframe. The data was extracted form a database used for tracking system troubles, called Remedy, and is currently stored in a mySQL database and available for export to Excel format. There are also some basic query and graph capabilities available. For more information on the data, please contact the PDSI researcher at NERSC: Akbar Mokhtarani or the Principal Investigator for PDSI at NERSC: John Shalf .

Data for Storage System failure

Data for Network Outages

NERSC Global File System failure and statistics

NERSC Systems Diskfailure

I/O Benchmarking and Characterization

Our goal is two-fold. First, we wish to serve as an information source for those working on I/O benchmarks or characterization. Second, we will select one or more petascale applications to do I/O benchmarking on. The results and information will be presented here.

I/O Benchmarking


Application Benchmarks

A compilation of performance benchmarks on scientific applications used at NERSC. We provide profiles of these applications using IPM and system calls trace.

Application Benchmarks