Hi Aaron, I use Sympy to compute the solution of an inverse problem (Ax = b) symbolically, where A is a 9-by-9 symbolic matrix. The computation is quite fast, however the out-printing of the solution ''x'' vector is very slow, even the first entry x[0] takes longer than 2 minutes. And the sympy.ratsimp(x[0]) is also very slow.. I am confused...
Hope for your kind response. Thanks in advance! Hong 在 2012年3月4日星期日UTC+1下午9时49分06秒，Aaron Meurer写道： > > On Wed, Feb 29, 2012 at 9:28 AM, Vinzenz Bargsten <[email protected] > <javascript:>> wrote: > > Am 28.02.2012 06:41, schrieb Ondřej Čertík: > > > >> Hi Vinzenz, > >> > >> On Mon, Feb 27, 2012 at 4:27 PM, Vinzenz Bargsten<[email protected] > <javascript:>> > >> wrote: > >> [...] > >>> > >>> Thank you very much. I think I get the point, but I will need some time > >>> to > >>> implement it. > >>> If this is working, I can finally get rid of Mathematica and the > complete > >>> modeling procedure as well as the required model transformations > >>> for a 7-axes robot (Kuka LWR) are carried out by sympy. > >> > >> That'd be really cool. Yes, the trigonometric simplification is pretty > >> hard. > >> Besides that, how is SymPy doing otherwise in terms of speed compared > >> to Mathematica for your problem? > >> > >> > >> Ondrej > > > > Hi Ondrej, > > > > your question is not easy to answer. It depends on the actual task and > how > > you implement it / the algorithm. The simplification problem is an > example: > > Using the expand()-replace()-solution is probably slower compared to > > Mathematica's Simplify[], but using the monomial-based solution may be > > faster. > > On average, I would say sympy is at least as fast as Mathematica for my > > application. > > Well, in general, you should be able to just call simplify() (or > trigsimp()), and it should just work. So simplification (especially > trig simplification) is one area where we can definitely use > improvement. > > > > > The advantages of sympy in my eyes compared to Mathematica are that > sympy is > > deterministic and you get error messages and the messages point to the > > problem. > > This is the first time I've heard this particular reason for SymPy > being better. How is Mathematica non-deterministic? I've never used > it, except for WolframAlpha, because, as you noted, it's not free, and > I don't own a copy. > > I suppose the better error messages are a result of Python including > the traceback? > > > Besides the math functions (in many cases superior or unique) > > Cool. What functions are unique? I only know of one that I'm pretty > sure is unique (or at least I couldn't find it in their docs), but > it's part of the polys module. I guess maybe the code generation > stuff? > > > it > > has many convenient functions (such as the printing functions to generate > > c-code of your expressions) and one can use all other python functions. > -- > > and sympy is free ;) > > > > Regards, > > Vinzenz > > Aaron Meurer > > -- 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/65cd2d04-84ff-4eac-8a69-93c144061d83%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
