[Nauty-list] Sparse or dense?

[Hans Georg Schaathun]

Hi,

I am using nauty for the equivalence of linear codes every now and
then.  Ten years ago I settled for the dense data structure.  Revisiting
the problem now, I note that there is a sparse data structure as well.

Is there any guidance available on when each data structure is optimal?
Or possibly performance comparisons?  I have tried searching without 
finding anything.

In my particular case, I have a bipartite graph with about 50-75
vertices in one set at a thousand in the other.  The vertices in
the larger set have degree 15-20.

Thanks in advance for any pointers or advice,
-- 
:-- Hans Georg