A summary of the results of the limited analysis of FORTRAN efficiency performed for this paper are presented below:

• An average memory penalty of 17% and time penalty of 16% was incurred for four types of processing (Table 2) in a Harris FORTRAN vs assembly language efficiency comparison. This was further confirmed by the combined benchmark totals for the Harris SLASH 4 (Table 5) where the memory penalty was 19% and the time penalty was 16%. According to Trainor and Burlakoff⁴, these results are consistent with published results for other comparable HOLS that have indicated a 10% to 20% penalty.

• An average memory savings of 43% and time savings of 29% (Table 3) was achieved in an optimizing vs. non-optimizing FORTRAN compiler comparison.

• A memory savings of 32% and time savings of 23% was achieved in a comparison of two optimizing compilers.

• Benchmark results for seven candidate simulation computers were presented exhibiting a wide range of memory and execution time efficiencies.

Menu

Frank Sigmund