Transforming Fortran DO loops to improve performance on vector architectures

Technical Report · Thu May 01 00:00:00 EDT 1986

OSTI ID:5607018

Cowell, W R; Thompson, C P

The performance of programs executing on vector computers is significantly improved when the number of accesses to memory can be reduced. Unrolling Fortran DO loops, followed by substitutions and eliminations in the unrolled code, can reduce the number of loads and stores. The unrolling transformation and associated transformations of Fortran DO loops are characterized, and a set of software tools to carry out these transformations is described. The tools use the machinery available in Toolpack and have been integrated into that environment. The results of applying these tools to a collection of linear algebra subroutines are included. 7 refs., 2 figs.

• Research Organization: Argonne National Lab., IL (USA)
• DOE Contract Number: W-31-109-ENG-38
• OSTI ID: 5607018
• Report Number(s): ANL-85-63; ON: DE86012086
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
VECTOR PROCESSING
ALGORITHMS
CRAY COMPUTERS
FORTRAN
PERFORMANCE
PROGRAMMING
SUPERCOMPUTERS
COMPUTERS
DIGITAL COMPUTERS
MATHEMATICAL LOGIC
PROGRAMMING LANGUAGES
990200* - Mathematics & Computers