Latency Reduction in Software-DSMs by Means of Dynamic Function Splicing

M. Klemm, R. Veldema, and M. Philippsen (Germany)

Keywords

Jackal, Java, DSM, Automatic Compiler Optimization

Abstract

We present a compilerbased technique to automatically identify and extract Remote Procedure Calls, socalled Function Splices, out of potentially arbitrary sequences of Java code compiled for a software DSM. The goal is to lower communication latencies and message traffic by re placing data shipping by function shipping. Dynamic Func tion Splicing dynamically decides at runtime whether to in voke a function splice on the local machine or to execute it remotely on the home node of the requested data. On proofofconcept microbenchmarks Dynamic Function Splicing reduces the execution wall time by ap proximately 29 %; about 25 % of the messages can be saved.

Important Links:



Go Back