Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Implementing fracture mechanics analysis on a distributed-memory parallel processor

Technical Report ·
OSTI ID:6505994
As part of an exploratory studies initiative at Oak Ridge National Laboratory, a parallel algorithm was developed and implemented for finite-element fracture mechanics. It was actually implemented on an Intel iPSC hypercube, although the algorithm was designed for any distributed-memory parallel computer with message passing primitives. A p-frontal method was developed for the solution of the equilibrium equations. This method required only log/sub 2/p communication exchanges between processors during the solution. On four processors, the parallel code solved a sample problem 2.7 times faster than the serial code, ORVIRT.PC, from which it was derived. 15 refs.
Research Organization:
Oak Ridge National Lab., TN (USA)
DOE Contract Number:
AC05-84OR21400
OSTI ID:
6505994
Report Number(s):
ORNL/TM-10367; ON: DE87009778
Country of Publication:
United States
Language:
English