- taskvine
- sade
- floability
- pledge
- xgfabric
- highlights
•
•
•
•
•
-
TaskVine Insights: DaskVine Executor for Practical Scientific Graphs
In the TaskVine world, workflows can be built task by task or as full graphs. This post focuses on DaskVine and shows how to run Dask graphs on TaskVine with practical guidance on execution modes, data movement, and scheduling options.
-
Pledge Development Updates
Our Pledge project continues to evolve, uncovering hidden dependencies in HPC scripts and making scalable workflows increasingly within reach.
-
TaskVine Insights: Picking the Right Task Surface in Python (Task, PythonTask, FunctionCall)
In the TaskVine API, Task, PythonTask, and FunctionCall are three ways to describe work. This post explains how each one expresses dependencies with File objects, and why libraries help avoid Python cold starts and reuse imports and global state.
-
Floability Winter Research Retreat 2026
Snow kept falling as we hosted our Floability winter retreat at Notre Dame, catching up with friends from UIUC, DePaul, and Missouri and hashing out some tough workflow problems together.
-
TaskVine Insights - Intermediate Data Model
Large DAGs often bottleneck on intermediate data. Here's how TaskVine's temp files keep intermediates on workers and cut manager-side transfers.