Quantcast
Viewing all articles
Browse latest Browse all 2210

Characterising pipeline workloads

Image may be NSFW.
Clik here to view.
© R&D Compute group

The image shows a possible analysis of the resource usage by pipelines. Understanding how pipelines affect the underlying hardware in a uniform way can for example be used to measure improvement of new pipelines or efficiently do observation scheduling.

The top-left image is a graphical representation of the pipeline we used (kindly provided by Marco Iacobelli). Underneath, a plot shows the usage of CPU and memory. The image of the right shows how much time is spent in each part of the pipeline in a hierarchical way.

The work has been presented at the Astronomical Data Analysis Software and Systems (ADASS) conference. A PDF version of the poster can be found on the website of the conference. The presented results, which will also appear in the conference proceedings, can be found on arXiv.


Viewing all articles
Browse latest Browse all 2210

Trending Articles