h5part AT lists.psi.ch
Subject: H5Part development and discussion
- From: Achim Gsell <Achim.Gsell AT psi.ch>
- To: Andreas Adelmann <andreas.adelmann AT psi.ch>
- Cc: Mark Howison <MHowison AT lbl.gov>, h5part AT lists.psi.ch, John Shalf <JShalf AT lbl.gov>
- Subject: Re: [H5part] Re: H5Block and ghost zones
- Date: Thu, 15 Jul 2010 15:02:10 +0200
- List-archive: <https://lists.web.psi.ch/pipermail/h5part/>
- List-id: H5Part development and discussion <h5part.lists.psi.ch>
On Jul 14, 2010, at 1:20 AM, Andreas Adelmann wrote:
> Hi Mark ok now I understand.
> In case of a stencil based computation with a very anisotropic computation
> domain I can think of such cases. So we probable want to keep this feature
> but maybe have a configure option to ride the more easy way; i.e. all the
> dimensions have the same number of ghost nodes. This just in case of you
> very well observed problem with O(1E6) cores.
> What do you think ?
After defining the layout, we don't need the information for all processes
any more. So, why not allocating the memory in the layout-defining function
and releasing it before we leave this function?
- [H5part] H5Block and ghost zones, Mark Howison, 07/08/2010
- [H5part] Re: H5Block and ghost zones, Mark Howison, 07/08/2010
- [H5part] Re: H5Block and ghost zones, Achim Gsell, 07/09/2010
- [H5part] Re: H5Block and ghost zones, Mark Howison, 07/09/2010
- Re: [H5part] Re: H5Block and ghost zones, Andreas Adelmann, 07/10/2010
Archive powered by MHonArc 2.6.19.