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

Automatic differentiation of a parallel molecular dynamics application

Conference ·
OSTI ID:555543
 [1]; ;  [2]
  1. Univ. of Illinois, Urbana, IL (US). Dept. of Computer Science
  2. Argonne National Lab., IL (US). Mathematics and Computer Science Div.
The ADIC and ADIFOR automatic differentiation tools have proven useful for obtaining the derivatives needed in many scientific applications written in Fortran 77 or ANSI C. But many new scientific programs are written for or ported to parallel platforms to achieve maximal performance. The authors provide an overview of the approach to the complex task of applying automatic differentiation techniques to parallel programming environments, especially as applied to a parallel molecular dynamics application written in C++ with PVM message passing.
Research Organization:
Argonne National Lab., IL (US)
Sponsoring Organization:
USDOE Office of Energy Research, Washington, DC (US); National Science Foundation, Washington, DC (US)
DOE Contract Number:
W-31109-ENG-38
OSTI ID:
555543
Report Number(s):
ANL/MCS-P--637-0197; CONF-970342--; ON: DE98001330; CNN: Cooperative Agreement CCR-9120008
Country of Publication:
United States
Language:
English