El El jue, 2 feb 2017 a las 5:24, Arne Babenhauserheide <arne_...@web.de> escribió:
> > Germán Diago <germandi...@gmail.com> writes: > > >> Aside from using a more recent Guile, I do not see obvious > >> optimizations, however (more exactly: all my tries to speedup the code > >> only made it slower). > > > > I was struggling with using read-line! instead of read-line but > > experimenting in the command line > > I cannot figure out how it works actually. read-line! will reuse a > buffered > > string. There was lots of time spent there, I am not sure if because of > > allocations or just because of I/O. If someone could tell me how to do > it... > > One thing I realized now is that you might be able to avoid the > string->number conversion completely since you’re outputting as text > anyway. That might save 1/4th of the time (but I did not experiment with > that now). True. > Also I’m not sure whether the "is this a number"-check could be sped up > by adding specialization for chars in the Scheme-compiler, similar to > how floating point comparisons were added in the latest release: > > https://git.dthompson.us/guile.git/commitdiff/79fa147990021b48a5ede2f33441bc95721fe63d > > > Thanks for your time Arne! > > Thank you for the challenging problem Welcome!! :)
