SYZYGY - A Framework for Scalable Cross-Module IPO

Performing analysis across module boundariesfor an entire program is important for exploitingseveral runtime performance opportunities. However,due to scalability problems in existing full-programanalysis frameworks, such performance opportunitiesare only realized by paying tremendous compile-timeco...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Moon, Sungdo, Li, Xinliang D., Hundt, Robert, Chakrabarti, Dhruva R., Lozano, Luis A., Srinivasan, Uma, Liu, Shin-Ming
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Performing analysis across module boundariesfor an entire program is important for exploitingseveral runtime performance opportunities. However,due to scalability problems in existing full-programanalysis frameworks, such performance opportunitiesare only realized by paying tremendous compile-timecosts. Alternative solutions, such as partialcompilations or user assertions, are complicated orunsafe and as a result, not many commercialapplications are compiled today with cross-moduleoptimizations.This paper presents SYZYGY, a practicalframework for performing efficient, scalable,interprocedural optimizations. The framework isimplemented in the HP-UX Itanium® compilers andwe have successfully compiled many very largeapplications consisting of millions of lines of code. Weachieved performance improvements of up to 40%over optimization level two and compilation timeimprovements in the order of 100% and morecompared to a previous approach.
DOI:10.5555/977395.977668