Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[web] Marketing / features index page #2123

Merged
merged 13 commits into from Jun 1, 2018

Conversation

Projects
None yet
5 participants
@bilke
Copy link
Member

bilke commented May 15, 2018

In preparation of the shutdown of the existing opengeosys.org page and as a todo from the usability initiative here is the first draft of a features-at-a-glance index page.

Check the page here!

Now I need your help: Please contribute feature descriptions and fancy images from your point of view! A few ideas:

  • Implemented processes worth mentioning, all @ufz/members
  • Interfaces to other data formats / applications
  • HPC and parallelization, @wenqing @TomFischer
  • pre-, post-processing, Data Explorer, @TomFischer @rinkk and me
  • Visualization, @rinkk and me
  • Python bindings? @chleh
  • Applications / projects, all
  • QA, benchmarking, me?
  • ... (more ideas from you)
  • Coupling schemes, @endJunction @wenqing
  • Software engineering, me
  • Impressum

Feel free to directly add to this PR! The pages content is at web/content/_index.pandoc. Or you can also post your description here as a comment and I will integrate it.

This PR should be finished on 1st of July latest.

@bilke bilke added the WIP 🏗 label May 15, 2018

@wenqing

This comment has been minimized.

Copy link
Member

wenqing commented May 16, 2018

@TomFischer Any comments? Maybe you have better figures.

High performance computing
High performance computing has became a powerful tool in the modelling of environmental and geotechnical problems for better characterization of the complexity of geo-systems as well
as predicting their evolution in time. The parallel computing is the most efficient method in the high performance computing. In the ogs, the parallalization of the finite element (FE) computation is based on the domain decomposition method (DDC).
ddc
In the parallel FEM of ogs, the decomposed global matricies and vectors are handled by the PETSc routines of matrix and vectors, and the system of linear equations are solved by PETSc solver. That means MPI is used in the parallel FE of ogs for general parallel computing architectures.

So far, the parallel FE of ogs works for the problems with single process, problems with coupled processes and using the same order of element for each process.

speed_3d

@wenqing

This comment has been minimized.

Copy link
Member

wenqing commented May 16, 2018

@endJunction Any comment?
Coupling Scheme

A coupled system of equations can be either solved in a fully coupled way of the monolithic method, or in the sequential manner of the staggered scheme (as shown in the following figure).
staggered
In current ogs6, the monolithic scheme is applied for all coupled processes, while the staggered scheme are available for the coupled processes of themo-hydraulic, hydro
-mechanical, and phase field mechanical problems.

@bilke I posted two descriptions as comment.

@TomFischer

This comment has been minimized.

Copy link
Member

TomFischer commented May 16, 2018

@wenqing , @bilke
I could create some figures from the fractured cube. I have alreade run time figures and scaling curves for different architectures (EVE, mistral, TAURUS and JURECA). But I believe this is nothing for the start page of the website.

bilke added some commits May 16, 2018

[web] Added gallery implementation.
Uses fancybox, thumbnails are automatically generated with sharp-loader.

Decomposed global matricies and vectors are handled by PETSc and the system of linear equations are solved by the performant PETSc solver. PETSc builds upon the Message Passing Interface (MPI) suitable for a wide variety of parallel computing architectures.

Parallelization is implemented for single processes as well problems with coupled processes which are using the same order of element for each process.

This comment has been minimized.

@wenqing

wenqing May 17, 2018

Member

single processes to single process`

@chleh

This comment has been minimized.

Copy link
Collaborator

chleh commented May 17, 2018

I just started working on some BCs based on Python. Once that's finished I can add some bullet points about it.

bilke added some commits May 17, 2018

[web] Added carousel galleries and vis and data int. drafts.
Carousel is implemented with slick-carousel and fancybox.

Also added a quicklink nav bar.

@bilke bilke force-pushed the bilke:web-start branch from 519562e to 345ce13 May 29, 2018

@bilke bilke force-pushed the bilke:web-start branch from 345ce13 to 8b6d0da May 29, 2018

bilke added some commits May 31, 2018

[web] Final changes for first release of marketing page.
Reorderd features a bit.
Uploaded precompiled JS and CSS.

@bilke bilke removed the WIP 🏗 label Jun 1, 2018

@bilke bilke merged commit f9db47a into ufz:master Jun 1, 2018

2 checks passed

continuous-integration/appveyor/pr AppVeyor build succeeded
Details
continuous-integration/jenkins/pr-merge This commit looks good
Details

@bilke bilke deleted the bilke:web-start branch Jun 1, 2018

@bilke

This comment has been minimized.

Copy link
Member Author

bilke commented Jun 1, 2018

@ufz/members

Is now online at www.opengeosys.org.

Old adresses such as docs.opengeosys.org, tutorials.opengeosys.org and benchmarks.opengeosys.org should redirect accordingly.

Future feature highlights are appreciated!

@waltherm

This comment has been minimized.

Copy link
Contributor

waltherm commented Jun 4, 2018

Hi all,

thanks for the new page layout. I would like to give some comments:

"Create meshes approximating geometrically the domain of interest. Analyze mesh quality, cleanup the mesh or adding layers to it." ->
"Create meshes fitting your domain of interest. Analyze mesh quality, cleanup the mesh or add layers to it."

"The monolithic scheme is applied for all coupled processes...". This description of the Process Coupling is slightly misleading. It is not the case that we can simply couple any process monolithically or iteratively. Maybe, we should reformulate that as "We provide monolithic or staggered coupled implementation following the requirements of the respective processes."?

Add a link to the OGS Data Explorer?

I would shift the HPC block right after Process Coupling.

Add a link to PETSc?

Add link to GitHub in the "Transparent Development Workflows" block?

Button for "Download Opengeosys" is broken.

Should we also add a link to the last release of OGS5?

Cheers,
Marc

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.