[GiNaC-devel] Re: [GiNaC-cvs] ginac flags.h mul.cpp power.cpp

[Richard B. Kreckel]

CVS Service wrote:
> Update of /home/cvs/GiNaC/ginac by vollinga
> 
> Modified Files:
>       Tag: ginac_1-4
> 	flags.h mul.cpp power.cpp 
> Log Message:
> Synced to HEAD:
> - This helps mul::expand() and friends to recognize objects which have
> no indices at all [Sheplyakov].

Chris,

We are curious what can be done about the performance regression one of 
your patches for dummy index renaming introduced. The seemingly 
innocuous patch has been identified by Jens and Alexei: 
<http://www.ginac.de/pipermail/ginac-list/2007-September/001215.html>.

At first, the above log message suggested to me that the problem has 
been dealt with. But that is not so: The program is still extremely 
slow. In fact, Alexei's patches are entirely unrelated. Since neither an 
explanation of what's happening nor a fix has been suggested, we would 
really appreciate some expert insight by you.

Thanks in advance
     -richy.
-- 
Richard B. Kreckel
<http://www.ginac.de/~kreckel/>