``Algorithm-Based Diskless Checkpointing for Fault Tolerant Matrix Operations''

James S. Plank, Youngbae Kim, and Jack Dongarra

25th International Symposium on Fault-Tolerant Computing, Pasadena, CA, June, 1995, pages 351-360.

Available via anonymous ftp to cs.utk.edu in pub/plank/papers/FTCS25.1995.ps and pub/plank/papers/FTCS25.1995.pdf.

The successor to this paper (to appear in JPDC in September, 1997) is here.

Abstract

This paper is an exploration of diskless checkpointing for distributed scientific computations. With the widespread use of the ``Network Of Workstation'' (NOW) platform for distributed computing, long-running scientific computations need to tolerate the changing and often faulty nature of NOW environments. We present high-performance implementations of several algorithms for distributed scientific computing, including Cholesky factorization, LU factorization, QR factorization, and preconditioned conjugate gradient. These implementations are able to run on PVM networks of at least $N$ processors, and can complete with low overhead as long as any $N$ processors remain functional. We discuss the details of how the algorithms are tuned for fault-tolerance, and present the performance results on a PVM network of SUN workstations, and on the IBM SP2.

Postscript of the paper

PDF of the paper


Citation Information