Practical Prefetching Techniques for Parallel File Systems

Adaptive Dual-Cache Scheme with Dynamic Prefetching Scheme in Parallel File System

An adaptive Dual-Cache Scheme(DCS) and dynamic prefetching scheme were designed to improve the performance of the Parallel File System for Linux(PFSL). PFSL is a parallel file system for a clustering environment and is implemented using a multi-threaded programming technique with POSIX thread libraries supported by Linux. The proposed adaptive DCS and dynamic prefetching scheme both reflect the...

A Status Report on Research in Transparent Informed Prefetching (CMU-CS-93-113)

This paper focuses on extending the power of caching and prefetching to reduce file read latencies by exploiting application level hints about future I/O accesses. We argue that systems that disclose high-level knowledge can transfer optimization information across module boundaries in a manner consistent with sound software engineering principles. Such Transparent Informed Prefetching (TIP) sy...

Prefetching in File Systems for MIMD Multiprocessors

Implementation and Evaluation of Prefetching in the Intel Paragon Parallel File System

The significant dijjerence between the speeds of the II0 system (e.g. disks) and compute processors in parallel systems creates a bottlerleck that lowers thepe$ormance of an applicatiorl that does a consideruble amount of disk accesses. A major portion of the compute processors’ time is wasted on wuitiqfor II0 to complete. This problem call be addressed ro a crrtairl. extent, if the Ilecessary ...

Path-Based Target Prediction for File System Prefetching

Prefetching is a well-known technique for mitigating the von Neumann bottleneck. In its most rudimentary form, prefetching simplifies to sequential lookahead. Unfortunately, large classes of applications exhibit file access patterns that are not amenable to sequential prefetching. More general purpose approaches often use models to develop an appropriate prefetching strategy. Such models tend t...