rytam2
commented
1 week ago Re: Pythia Cookoff This is my first time attending a cookoff - which I think is a great first cookoff experience. The week has been very productive, even though I missed out the intros of projects on day1, but I didn't miss out much from attending the break out session later that day. Within my group, we have a pretty clear distribution of work, which I really appreciate, as I get to focus on what I have to work on throughout the week (mainly writing and doing lit review on radiative feedbacks).
On the Pythia community as a whole, I love that the community is very supportive (Love the google meet function - now I prefer Google meet wayyyyyyy over Zoom for the explosive reactions). Also getting to meet others who have the same specific interests as mine on climate feedbacks is super fun; meeting others working in the climate science field but on other topics like paleoclimate - it's very cool to learn about what important questions people are trying to ask, and seeing them also struggling with data is very reassuring; and definitely meeting engineers who develop the tools - eye-opening and excited to use these new tools to check out my research results! And seeing all the work that engineers are doing on developing definitely adds more to my already-high respect and appreciation to them, and feel like more of it should be given from the community.
and grateful for the week-long free coffee and seeing Philip in-person :)))
Week 5
Overview
Meetings & Important Dates
Pythia Cookoff Follow-up
In the comments below, feel free to comment on anything you learned, found helpful, or enjoy from the Pythia Cookoff.
HPC Deliverable
Topological Aggregations
The first set of functionality you will be working on will be Topological Aggregations. Philip has put together an initial implementation of 2/6 of the types of aggregations. These implementations only use NumPy and the goal is to explore the use of Dask, CuPy and other packages to improve the performance.
Gradients
TODO