jim@THRUSH.STANFORD.EDU (Jim Helman) (04/12/88)
> From: riedesel@eliot.cs.uiuc.edu > > Gabriel Benchmark Suite tested on: > IBM RT PC > Explorer 2 > Explorer 1 > SUN 3/160 > (all times are in seconds) Thanks, but...... Could you, and any additional posters of benchmark timings, please give a few more details, such as which lisp you are running and what version of it, e.g. Lucid Common Lisp 2.1? Also please specify any additonal hardware or compiler directives, e.g. (compiler-options :target '68020/68881), the speed and saftey levels used and how extensively type declarations were made and tailored to the machine. > Disclaimer: All times are guaranteed to be approximate. Guarantee confirmed. The correspondence between the recently posted results with the Sun-3/160 numbers given with a July 1986 document from Sun "Sun Common Lisp Performance Report", is not good. Some of the Sun numbers are 2x slower and some others are 2x faster (TRIANG is 4X faster) than the recent posting. I know for a fact that the current Sun/Lucid 2.1 release is much better than the old Sun/Lucid release, which could account for the old Sun numbers being slow. Wisely and lucidly made declarations probably account for the cases where the old Sun numbers are faster. It's tough enough as it is to translate benchmark performance into real performance without adding additional unknowns. I wish people from companies such as Lucid, Franz, TI, Sun, Symbolics, etc., or those who have copies of their recent lisp performance reports, would post some good, optimized benchmark results. Thanks... Jim Helman Department of Applied Physics Stanford University