H5Part development and discussion

["Adelmann Andreas (PSI)" <andreas.adelmann AT psi.ch>]

ok lets do Hangouts, I will send you an invitation

A

------
Dr. sc. math. Andreas (Andy) Adelmann
Staff Scientist
Paul Scherrer Institut WBBA/219 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Phone Home: xx41 62 891 91 44
-------------------------------------------------------
Friday: ETH CAB H 85.1 or HPK G 28  +41 44 632 75 22
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 01 Nov 2016, at 14:47, Jozsef Bakosi <jbakosi AT lanl.gov> wrote:

Hi Achim,

Sorry for the late reply.

Yes, Wednesday 8am MST (5pm CET) works for me. I prefer Google hangout
(jbakosi AT gmail.com), but Skype might also work (jbakosi).

Let's have a chat on Wednesday then.

Thanks,
Jozsef
--
Jozsef Bakosi
Computational Physics and Algorithms (CCS-2)
MS D413, Los Alamos National Laboratory
Los Alamos, NM 87545, USA
email: jbakosi AT lanl.gov
phone: 505-665-0950
fax: 505-665-4972

On 10.28.2016 16:21, gsell wrote:

Hi Jozsef,

On 25 Oct 2016, at 21:50, Jozsef Bakosi <jbakosi AT lanl.gov> wrote:

Hi Andy,

Yes, I have seen h5hut but since at the time I was only interested in particles,
I just used H5Part. However, now I would be interested in it more and looked at
it again and saw that there is capability in there to save unstructured meshes
too. This would be great, if it worked in parallel. According to the web site it
does not yet work in parallel. I also have not found documentation on the API
for H5hut. I would be interested in exploring how much work is missing to make
it fulfill the requirements I am aiming for.

An experimental code for reading unstructured meshes in parallel already exist. But there is still lot of work to do, till this code is really usable. For the time being we cannot store unstructured meshes per time step, but this would be a minor change.

1. Does anyone know of an existing library that satisfies all the above
requirements?

No.

2. How much work do you think it would be to add the mesh topology information
to H5Part?

In principal this is already possible with H5hut/H5Part. It's not really according the data model for storing particle data, but anyway ...

A better solution would be to implement another data model for arbitrary datasets. This shouldn't be to hard.

I am available for live conversation, possibly via skype, google hangout, or
phone. Please let me know if someone is interested in pursuing this and want to
have a live discussion on it.

OK, let's have a live discussion on it. Would next Wednesday at 5pm CET (8am MST?) be OK?

Best

Achim