You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I personally used a docker image for running 100,000-millions of sims on the Open Science Grid. It's convenient (or maybe they require?) singularity images to run on OSG. It also needs to be lightweight. millions of distributed sims will get kind of heavy if you have to lug around 2.5G instead of 500MB.
Like, I said earlier, I'd vote each tskit tool having it's own light weight docker image (Ubuntu base), and then there can be a bigger tskit ecosystem, with all of the tools on top of a data science base. For example, I made this image for myself to have all the tools I wanted https://hub.docker.com/r/agladstein/data_science_popgen_notebook. We could make an tskit official data_science_tskit
Sounds good to me @agladstein - I don't really have much of an opinion, as I don't really use Docker images except to create JupyterHub instances for training.
It would be nice to have Binder working, though (#1413)
I personally used a docker image for running 100,000-millions of sims on the Open Science Grid. It's convenient (or maybe they require?) singularity images to run on OSG. It also needs to be lightweight. millions of distributed sims will get kind of heavy if you have to lug around 2.5G instead of 500MB.
Like, I said earlier, I'd vote each tskit tool having it's own light weight docker image (Ubuntu base), and then there can be a bigger tskit ecosystem, with all of the tools on top of a data science base. For example, I made this image for myself to have all the tools I wanted https://hub.docker.com/r/agladstein/data_science_popgen_notebook. We could make an tskit official data_science_tskit
Anyone else have an opinion? @jeromekelleher
Related to #1566
The text was updated successfully, but these errors were encountered: