> Thorsten Glebe wrote:
> ... snip ...
> >
> > All these problems do not exist of course with trivial Vector/Matrix
> > implementations like those of the CLHEP library. Those
> > you can easily incorporate in CINT, but as I mentioned, the performance is
> > (even in compiled code) low.
> >
> ... snip...
>
> I guess it is a matter of my complete ignorance in these issues, but could
> somebody on the thread elaborate a bit on why C++ implementations of matrix
> operations are supposed to suffer from low performance? What causes the C++
> code doing the matrix inversion to be slow compared to its FORTRAN equivalent?
I did not investigate one's packages performance, but from my own experience calling
virtual functions takes a lot of time. Sometimes it is much faster to replace it with the plain
"switch /case" .
Second: the "virtual function" does not allow to use "in-line" optimization.
There are other reasons as well.
Valeri
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:50:46 MET