Automatic Monitoring of Memory Hierarchies in Threaded Applications with AMEBA

* Final gross prices may vary according to local VAT.

Get Access

Abstract

In this paper we present an approach to online automatic monitoring of memory hierarchies in threaded applications. Our environment consists of a monitoring system and an automatic performance analysis tool. The CMM monitoring system uses static instrumentation of the source code and information from the hardware counters to generate performance data for selected code regions and data structures. The monitor supports threaded applications by providing per-thread performance data or by aggregating it. It also provides a monitoring request API for the performance tools. Our tool AMEBA performs an online automatic search for cache and thread-related ASL properties in the code.

This work is part of the EP-CACHE project funded by the German Federal Ministry of Education and Research (BMBF) and the Periscope project funded by the German Science Foundation (DFG).