Hi Simon and Martin,
I know Strassen-Winograd. Strassen uses 18 additions, Strassen-Winograd only
15, which is optimal for 7 multiplications ;-)
With the in-memory variant I mean: "Boyer, Dumans, Pernet and Zhou: Memory efficient
scheduling of Strassen-Winograd's matrix multiplication algorith
On Tuesday 12 July 2011, Volker Braun wrote:
> Does linbox actually use blas for fast matrix multiplication over finite
> fields? I know its hard to tell from the linbox source code but I don't
> think floating point stuff could be faster than the direct application of
> Strassen's algorithm.
Yes.
Does linbox actually use blas for fast matrix multiplication over finite
fields? I know its hard to tell from the linbox source code but I don't
think floating point stuff could be faster than the direct application of
Strassen's algorithm.
--
To post to this group, send an email to sage-devel
> Is there a compelling reason why linbox multiplication is not the
> default for dense matrices over finite fields?
IIRC, this dates back to the time when we didn't ship ATLAS, i.e. LinBox could
have been built against some reference implementation BLAS (== slow).
It wasn't touched afterwards
