Am Mittwoch, den 17.02.2010, 23:34 -0500 schrieb Raymond Toy:
> Recompiling the world seems to have taken care of most issues. The
> testsuite takes a very long time to run, however. Ecl takes a good 45
> minutes on a 2 GHz macmini, and it seems a large fraction of the time is
> spent on rtest_integrate. rtest_gamma and rtest_expintegral also takes
> quite a bit of time.
>
> CMUCL takes 700 sec, which is reasonable. But I'm also seen some kind
> of intermittent error where $dependencies is somehow set to ((mlist
> simp) 0 . 0). Kill doesn't like this.
>
> I haven't tried clisp yet.
Yes, there might be the problem that the testsuite is slower. I have not
figured out the details, but I will investigate this point more
carefully.
One point is, that I have committed code to do more simplifications for
the abs function. The file rtest_integrate generates a lot of results
with the abs function. Therefore, the code to simplify the abs function
might cause problems. I have to check it more carefully.
Another problem of rtest_integrate is, that a lot of examples need a lot
of time because of the way the algorithm are implemented. Because I did
not want to change anything, the algorithm is called after all other
steps of the integration routines. At last, the pattern of the power
functions are checked. It is possible to implement this much more
efficient.
Dieter Kaiser