More on memq, a timing experiment



I confirmed that compiling simp and loading simp makes Maxima slower (at
least for expand).
I did this using 5.14.99rc1, GCL 2.6.8, Windows XP, and .... Before loading
simp.o, expand(h) took 1.856 seconds (average of five) and after loading
simp.o,
expand took 4.686 seconds :(


(%i2) h : (1+x+y+z+t)^25;

(%o2) (z+y+x+t+1)^25
(%i3) (expand(h), expand(h), expand(h), expand(h), expand(h))$
Evaluation took 9.2800 seconds (9.2800 elapsed)

(%i4) compile_file("C:/Program Files/Maxima-5.14.99rc1/share/maxima/5.14.99
rc1/src/simp.lisp")$

(%i5) load("C:/Program Files/Maxima-5.14.99rc1/share/maxima/5.14.99
rc1/src/simp.o")$

(%i6) (expand(h), expand(h), expand(h), expand(h), expand(h))$
Evaluation took 23.4300 seconds (23.4300 elapsed)


Barton