Omnisc'IO: A Grammar-Based Approach to Spatial and Temporal I/O Patterns Prediction

Matthieu Dorier 1, 2 Shadi Ibrahim 1 Gabriel Antoniu 1 Robert Ross 3
1 KerData - Scalable Storage for Clouds and Beyond
Inria Rennes – Bretagne Atlantique , IRISA-D1 - SYSTÈMES LARGE ÉCHELLE
3 MCS
ANL - Argonne National Laboratory [Lemont]
Abstract : The increasing gap between the computation performance of post-petascale machines and the performance of their I/O subsystem has motivated many I/O optimizations including prefetching, caching and scheduling techniques. To further improve these techniques, modeling and predicting spatial and temporal I/O patterns of HPC applications as they run has became crucial. In this paper we present Omnisc'IO, an approach that builds a grammar-based model of the I/O behavior of any HPC application, and uses it to predict when future I/O operations will occur, where and how much data will be accessed. Omnisc'IO is transparently integrated into the POSIX and MPI I/O stacks, and does not require any modification in applications or higher level I/O libraries. It works without any prior knowledge of the application, and converges towards accurate predictions within a couple of iterations only. Its implementation is very efficient both in computation time and memory footprint.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [30 references]  Display  Hide  Download

https://hal.inria.fr/hal-01025670
Contributor : Matthieu Dorier <>
Submitted on : Thursday, July 24, 2014 - 9:38:55 AM
Last modification on : Thursday, November 15, 2018 - 11:57:44 AM
Document(s) archivé(s) le : Monday, November 24, 2014 - 7:46:11 PM

File

paper.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01025670, version 1

Citation

Matthieu Dorier, Shadi Ibrahim, Gabriel Antoniu, Robert Ross. Omnisc'IO: A Grammar-Based Approach to Spatial and Temporal I/O Patterns Prediction. SC14 - International Conference for High Performance Computing, Networking, Storage and Analysis, IEEE, ACM, Nov 2014, New Orleans, United States. ⟨hal-01025670⟩

Share

Metrics

Record views

935

Files downloads

524