1 / 2

Enhancing Scalable Online Performance Measurement with Smon and Tau Integration

This document explores the integration of Smon control and Tau measurement for enhancing data processing capabilities in scalable computing environments. By intelligently managing what data to dump and measure, we can significantly reduce the volume of data processed. The paper discusses how aggregate topology and coarse-grained data management contribute to improved performance. Additionally, we highlight the advantages of real-time data handling over traditional file systems, with diagrams illustrating the overall architecture and internal coupling between Smon and Tau. Future efforts aim to prototype larger-scale implementations to improve performance metrics.

Télécharger la présentation

Enhancing Scalable Online Performance Measurement with Smon and Tau Integration

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. 8 7 2 1 • 1.Control using Smon: • Control what to dump • Control what to meas • Reverse s-exp chanel • So control can reduce • Amt of data dumpd Ex. Of Adding more Intelligence to the Transport To Make is scalable. (whats the adv. Of add To transport vs. tau) Show 3 ways to scale. Abstract and stuff • Rationale • Why need online • Why not file-sys • Why use cluster-m • etc 3 9 10 4 3.Scalable Topology and Data Aggregation Again reduce by Aggregating and Making data coarse Grained. Diagram of the Overall setup Diagram of Internal Tau/ Smon Coupling 2.Sampling over the Ranks in each time Step. Sampling also reduces Amt of data (but over Diff. dimension). 12 5 6 11 Conclude Future efforts Prototyping on Larger scale platforms Changes to Smon Changes to Tau Performance numbers Over NFS: W/o Tau, wTau, wDump Performance-imp factor Over Smon: Same cases as above Compare to MRNet Or Show setup on BGL

  2. 8 7 2 1 • 1.Control using Smon: • Control what to dump • Control what to meas • Reverse s-exp chanel • So control can reduce • Amt of data dumpd Ex. Of Adding more Intelligence to the Transport To Make is scalable. (whats the adv. Of add To transport vs. tau) Show 3 ways to scale. Abstract and stuff • Rationale • Why need online • Why not file-sys • Why use cluster-m • etc 3 9 10 4 3.Scalable Topology and Data Aggregation Again reduce by Aggregating and Making data coarse Grained. Diagram of the Overall setup Diagram of Internal Tau/ Smon Coupling 2.Sampling over the Ranks in each time Step. Sampling also reduces Amt of data (but over Diff. dimension). 12 5 6 11 Conclude Future efforts Prototyping on Larger scale platforms Changes to Smon Changes to Tau Performance numbers Over NFS: W/o Tau, wTau, wDump Performance-imp factor Over Smon: Same cases as above Compare to MRNet Or Show setup on BGL Scalable Online Performance Measurement over a Cluster Monitor TAU __________ Supermon TAU / UO LOGOS LOS ALAMOS ACL LOGO

More Related