Summary: A General Interprocedural Framework for Placement of
Split­phase Large Latency Operations \Lambda
Gagan Agrawal
Department of Computer and Information Sciences
University of Delaware
Newark DE 19716
302­831­2783
agrawal@cis.udel.edu
Abstract
Overlapping split­phase large latency operations with computations is a standard technique for
improving performance on modern architectures, In this paper, we present a general interprocedural
technique for overlapping such accesses with computation. We have developed an Interprocedural Bal­
anced Code Placement (IBCP) framework, which performs analysis on arbitrary recursive procedures
and arbitrary control flow and replaces synchronous operations with a balanced pair of asynchronous
operations. We have evaluated this scheme in the context of overlapping I/O operations with com­
putation. We demonstrate how this analysis is useful for applications which perform frequent and
large accesses to disks, including applications which snapshot or checkpoint their computations or
out­of­core applications.
1 Introduction
Modern architectures have deep memory hierarchies. Processors have one or two levels of cache, followed

Source: Agrawal, Gagan - Department of Computer Science and Engineering, Ohio State University

Collections: Computer Technologies and Information Sciences