The following forwarded from the CommonMusic mailing list - original
mail by Heinrich Taube. My congratulations to the SBCL developers:
-------------------------------------------------------------------------
now that cm runs in most cls and three schemes (port to STklos scheme is
not quite done...) Ive added a timing suite to the distro (etc/time.cm) so
I can keep track of whats going on. here are some results for your
amusement.
Ive lumped Schemes and CLs apart from each other. Stklos is in the
ballpark with CLISP, and gauche is not that far behind. I dont know why
Guile is losing so big, it may have to do with testing on a ppc instead of
x86, or maybe ive screwed up somewhere. but stklos and gauche do pretty
well!
MACHINE: Dual 1.8 GHz PPC/G5, 1GIG RAM, OS X 10.3.9
#1 reading cyclic pattern 10000 times
#2 reading random pattern 10000 times
#3 running process 10000 times
#4 outputting 10000 midi notes to nowhere
#5 appending 1000 midi notes to seq
#6 3 processes to midifile, each 1000 notes/random waits
#1 #2 #3 #4 #5 #6
Gauche 0.8.3 0.657 1.250 0.275 0.673 4.506 0.515
Guile 1.7.1 1.000 36.000 7.666 18.000 17.333 6.333
STklos 0.61 0.360 1.555 0.340 0.786 1.383 ?????
CLISP 2.33 0.269 0.396 0.170 0.365 0.803 0.427
OpenMCL 0.14.2 0.047 0.117 0.150 0.222 0.438 0.235
SBCL 0.8.17 0.030 0.041 0.013 0.630 0.224 0.271
-------------------------------------------------------------------------
Ralf Mattes
|