Evaluation of Compiler Optimizations for Fortran D on MIMD Distributed-Memory Machines (1992)
| Venue: | IN PROCEEDINGS OF THE 1992 ACM INTERNATIONAL CONFERENCE ON SUPERCOMPUTING |
| Citations: | 33 - 11 self |
BibTeX
@INPROCEEDINGS{Hiranandani92evaluationof,
author = {Seema Hiranandani and Ken Kennedy and Chau-Wen Tseng},
title = {Evaluation of Compiler Optimizations for Fortran D on MIMD Distributed-Memory Machines},
booktitle = {IN PROCEEDINGS OF THE 1992 ACM INTERNATIONAL CONFERENCE ON SUPERCOMPUTING},
year = {1992},
pages = {86--100},
publisher = {}
}
OpenURL
Abstract
The Fortran D compiler uses data decomposition specifications to automatically translate Fortran programs for execution on MIMD distributed-memory machines. This paper introduces and classifies a number of advanced optimizations needed to achieve acceptable performance; they are analyzed and empirically evaluated for stencil computations. Profitability formulas are derived for each optimization. Results show that exploiting parallelism for pipelined computations, reductions, and scans is vital. Message vectorization, collective communication, and efficient coarsegrain pipelining also significantly affect performance.







