[CM] TeXmacs & S7
bil at ccrma.Stanford.EDU
bil at ccrma.Stanford.EDU
Tue Jan 18 05:59:52 PST 2022
It's not the size of the environment that matters;
look at libgsl.scm (or the timing test tools/tgsl.scm) --
it creates the *libgsl* environment that has about 2600
entries, but lookup_from takes up maybe 2% of the total
time). Or tools/tlet.scm which creates an environment
with 1000000 entries, then sums them all and so on.
It essentially never has to do a linear search.
If I can find the time, I'll try to see what is
happening in texmacs. annotate_callgrind output
from a sample run might help. (Run texmacs
under valgrind --tool=callgrind, then run the output
of that through annotate_callgrind --auto=yes --show-percs=no
--threshold=100 callgrind.out.<xxxx> > output) where
<xxxx> is the number at the end of the callgrind output
file name.)
More information about the Cmdist
mailing list