dolfin team mailing list archive
-
dolfin team
-
Mailing list archive
-
Message #15460
Re: [HG DOLFIN] Update ChangeLog and test fix for missing hgbisect message from hg when
On Thursday 17 September 2009 19:22:49 DOLFIN wrote:
> One or more new changesets pushed to the primary dolfin repository.
> A short summary of the last three changesets is included below.
>
> changeset: 7089:86efd19527eb
> tag: tip
> user: Anders Logg <logg@xxxxxxxxx>
> date: Thu Sep 17 19:22:41 2009 +0200
> files: ChangeLog
> description:
> Update ChangeLog and test fix for missing hgbisect message from hg when
> pushing to fenics.org. (I just upgraded hg on the server.)
I like the new layout on the front end of www.fenics.org/hg/foo!
Nice with the branch view too.
Johan
>
> changeset: 7088:e5ee11714cf3
> user: Anders Logg <logg@xxxxxxxxx>
> date: Thu Sep 17 19:12:31 2009 +0200
> files: site-packages/dolfin/norm.py
> description:
> Add support for vectors in norm() function. Now handles both vectors
> and functions:
>
> norm(x)
> norm(v)
> norm(x, "l2")
> norm(v, "Hdiv")
>
>
> changeset: 7087:75562cd711a0
> user: "Johan Hake <hake@xxxxxxxxx>"
> date: Thu Sep 17 11:10:44 2009 +0200
> files: bench/fem/speedup/analyse-bench bench/fem/speedup/commit-bench
> description:
> Updates to parallell benchmark scripts
>
> First results:
>
> Speedup | Assemble Assemble + solve
> --------------------------------------
> 1 | 1 1
> 2 | 0.7 0.26568
> 4 | 0.43333 0.19166
> 8 | 0.26667 0.13381
> 16 | 0.33333 0.084034
> 32 | 0.1 0.10149
>
> This indicates that the direct serial solver is not that good,
> (is it possible to use MUMPS when in serial?)
> and that solve does not scale particular well.
>
> ----------------------------------------------------------------------
> For more details, visit http://www.fenics.org/hg/dolfin
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev
>
References
