I just looked briefly at the documentation. My guess is that the data structures and the organization (templates) are a major problem. I have no idea if the algorithms are inefficient. I think the system may also suffers from Greenspun's tenth rule. Except having chosen C++ instead of C to re-implement the Lisp-equivalent underpining, the inefficiency of the design affects almost everything.
My theory, anyway. RJF PS So far as I know, sympy is slower than the equivalent (free) Lisp code. So yes, it is re-inventing the wheel. Worse, it is reproducing the chain of errors in design that led to existing computer algebra systems. Or as we say, standing on the toes of giants. Look it up in wikipedia, esp. regarding Newton and Hook. On Sunday, April 27, 2014 6:06:46 AM UTC-7, Ondřej Čertík wrote: > > On Sun, Apr 27, 2014 at 6:38 AM, Joachim Durchholz > <[email protected]<javascript:>> > wrote: > > Am 26.04.2014 21:50, schrieb Ondřej Čertík: > > > >> Thanks Alan. Yes, I did. I tried simple benchmarks and SymbolicC++ is > even > >> slower than Sympy. > > > > > > Wow. > > So much for the assumption "C++ is generally faster than Python". > > At best, it's potentially faster :-) > > It is generally (a lot) faster, but you need to use the same > algorithms. Why apparently > SymbolicC++ doesn't use. > > Ondrej > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/sympy. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/966decf8-530d-4433-9ece-d1638f628234%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
