Swift machine learning model serving scheduling: a region based reinforcement learning approach

H Qin, S Zawad, Y Zhou, L Yang, D Zhao… - Proceedings of the …, 2019 - dl.acm.org
The success of machine learning has prospered Machine-Learning-as-a-Service (MLaaS)-
deploying trained machine learning (ML) models in cloud to provide low latency inference …

Clarisse: A middleware for data-staging coordination and control on large-scale hpc platforms

F Isaila, J Carretero, R Ross - 2016 16th IEEE/ACM …, 2016 - ieeexplore.ieee.org
On current large-scale HPC platforms the data path from compute nodes to final storage
passes through several networks interconnecting a distributed hierarchy of nodes serving as …

A cost-intelligent application-specific data layout scheme for parallel file systems

H Song, Y Yin, Y Chen, XH Sun - … of the 20th international symposium on …, 2011 - dl.acm.org
I/O data access is a recognized performance bottleneck of high-end computing. Several
commercial and research parallel file systems have been developed in recent years to ease …

Collective caching: Application-aware client-side file caching

W Liao, K Coloma, A Choudhary, L Ward… - … Symposium on High …, 2005 - ieeexplore.ieee.org
Parallel file subsystems in today's high-performance computers adopt many I/O optimization
strategies that were designed for distributed systems. These strategies, for instance client …

An implementation and evaluation of client-side file caching for MPI-IO

W Liao, A Ching, K Coloma… - 2007 IEEE …, 2007 - ieeexplore.ieee.org
Client-side file caching has long been recognized as a file system enhancement to reduce
the amount of data transfer between application processes and I/O servers. However …

Design and evaluation of multiple-level data staging for blue gene systems

F Isaila, J Blas, J Carretero, R Latham… - IEEE Transactions on …, 2010 - ieeexplore.ieee.org
Parallel applications currently suffer from a significant imbalance between computational
power and available I/O bandwidth. Additionally, the hierarchical organization of current …

View-based collective i/o for mpi-io

JG Blas, F Isaila, DE Singh… - 2008 Eighth IEEE …, 2008 - ieeexplore.ieee.org
This paper presents the design and implementation of a new file system independent
collective I/O optimization based on file views: view-based collective I/O. View-based …

I/o scheduling service for multi-application clusters

A Lebre, G Huard, Y Denneulin… - 2006 IEEE international …, 2006 - ieeexplore.ieee.org
Distributed applications, especially the ones being I/O intensive, often access the storage
subsystem in a nonsequential way (stride requests). Since such behaviours lower the …

Dynamic-CoMPI: Dynamic optimization techniques for MPI parallel applications

R Filgueira, J Carretero, DE Singh, A Calderon… - The Journal of …, 2012 - Springer
This work presents an optimization of MPI communications, called Dynamic-CoMPI, which
uses two techniques in order to reduce the impact of communications and non-contiguous …

A new flexible MPI collective I/O implementation

K Coloma, A Ching, A Choudhary… - 2006 IEEE …, 2006 - ieeexplore.ieee.org
The MPI-IO standard creates a huge opportunity to break out of the traditional file system I/O
methods. As a software layer between the user and the file system, an MPI-IO library can …