Dask: Parallel Computation with Blocked algorithms and Task Scheduling

Dask: Параллельные вычисления с использованием блочных алгоритмов и планирования задач
Matthew Rocklin
2015-01-01

DaskNumPyblocked algorithmsdynamic task schedulingout-of-core computation
Dask enables parallel and out-of-core computation. We couple blocked algorithms with dynamic and memory aware task scheduling to achieve a parallel and out-of-core NumPy clone. We show how this extends the effective scale of modern hardware to larger datasets and discuss how these ideas can be more broadly applied to other parallel collections.
1
Dask combines blocked algorithms with dynamic, memory-aware task scheduling for parallel and out-of-core computation.
2
Dask extends the effective scale of modern hardware to datasets larger than available memory.
3
The paper argues that blocked algorithms and memory-aware task scheduling can generalize to other parallel collection types.
4
The system provides a parallel, out-of-core computational framework that functions as a NumPy clone.

Dask's parallel and out-of-core computation system for blocked array algorithms

The scalability and execution efficiency achieved by coupling blocked algorithms with dynamic, memory-aware task scheduling

Publication Details
Publication Date
2015-01-01
Journal
Publisher
ISSN
Access Type
Author Information
Authors
Matthew Rocklin
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat
Make a presentation
100%