LISP and Symbolic Computation, 3(2)171-204
Endpaper: Performance of Two Common Lisp Programs on Several SystemsRichard Acuff, Stanford University, Stanford, California 94305
| Abstract: To assist in the evaluation of Lisp platforms for the Stanford University Knowledge Systems Laboratory, 25 Common Lisp implementations were benchmarked. Run time and compilation time data on two moderate-sized application programs are presented, along with data on the effect of compiler optimization levels and on the impact of display I/O on run time. For these Lisp benchmarks, several systems did not rank where we expected them based on speed ratings using other conventional measures. Also, the rankings of machines by Lisp speed differed for the two programs we tested. The data indicate that the performance of Lisp systems is very application dependent. The quality of the software environment should play at least as strong a role in machine selection as performance benchmarks. |
This article is not available online.
![[picture of journal cover]](../LISP.JPG)