Benchmarking the solar dynamo with Maxima



On Friday 28 November 2008 21:24:46 Adam Majewski wrote:
<skipped>
By the way, Prof Fateman gave me another reference for his old paper that 
shows how Macsyma was used for an approximate analytical solution of one-
dimensional alpha^2 Earth dynamo,  
http://www.cs.berkeley.edu/~fateman/papers/alpha-dynamo.pdf. 
> > compiled against the fastest lisps like sbcl or cmucl-sse2
Generally, gcl is ok. However it is not as good in memory management as sbcl 
and cmucl. So in the large problem its slow-down compare them is factor 10 or 
so. To test the  sophisticated solar dynamo model I have to use 300x300 
symbolic matrices. Gcl fails on this task.
> 1. Why Maxima uses slower version of lisp ?
Why? Some of the popular linux distributions like, fedora, gentoo and altlinux 
has sbcl compiled maxima rpms. There is sbcl for windows as well.
> 2. How to compile it against the fastest lisps ?
>From the sources, it is easy. Install the needed lisp. Then ./configure --
enable-sbcl(cmucl)

regards
V