Adapt or become extinct! The case for a unified framework for deployment-time optimization (position paper)

Goumas, G; McKee, SA; Sjalander, M; Gross, TR; Karlsson, S; Probst, CW; Zhang, L

dc.contributor.author	Goumas, G	en
dc.contributor.author	McKee, SA	en
dc.contributor.author	Sjalander, M	en
dc.contributor.author	Gross, TR	en
dc.contributor.author	Karlsson, S	en
dc.contributor.author	Probst, CW	en
dc.contributor.author	Zhang, L	en
dc.date.accessioned	2014-03-01T02:52:51Z
dc.date.available	2014-03-01T02:52:51Z
dc.date.issued	2011	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/36116
dc.subject	high-efficiency computing	en
dc.subject	runtime adaptation	en
dc.subject	specialization	en
dc.subject.other	Compile time	en
dc.subject.other	CPU architecture	en
dc.subject.other	Dynamic information	en
dc.subject.other	Execution context	en
dc.subject.other	Hardware characteristics	en
dc.subject.other	High efficiency	en
dc.subject.other	High-performance computing	en
dc.subject.other	Information flows	en
dc.subject.other	Just in time	en
dc.subject.other	Memory hierarchy	en
dc.subject.other	Memory organizations	en
dc.subject.other	Memory wall	en
dc.subject.other	ON dynamics	en
dc.subject.other	Performance Gain	en
dc.subject.other	Performance monitoring	en
dc.subject.other	Position papers	en
dc.subject.other	Power walls	en
dc.subject.other	Runtime adaptation	en
dc.subject.other	Seamless integration	en
dc.subject.other	specialization	en
dc.subject.other	Static information	en
dc.subject.other	Unified framework	en
dc.subject.other	User input	en
dc.subject.other	Computer programming languages	en
dc.subject.other	Computer software selection and evaluation	en
dc.subject.other	Memory architecture	en
dc.subject.other	Network architecture	en
dc.subject.other	Parallel programming	en
dc.subject.other	Software design	en
dc.subject.other	Static analysis	en
dc.subject.other	Computer systems programming	en
dc.title	Adapt or become extinct! The case for a unified framework for deployment-time optimization (position paper)	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1145/2000417.2000422	en
heal.identifier.secondary	http://dx.doi.org/10.1145/2000417.2000422	en
heal.publicationDate	2011	en
heal.abstract	The High-Performance Computing ecosystem consists of a large variety of execution platforms that demonstrate a wide diversity in hardware characteristics such as CPU architecture, memory organization, interconnection network, accelerators, etc. This environment also presents a number of hard boundaries (walls) for applications which limit software development (parallel programming wall), performance (memory wall, communication wall) and viability (power wall). The only way to survive in such a demanding environment is by adaptation. In this paper we discuss how dynamic information collected during the execution of an application can be utilized to adapt the execution context and may lead to performance gains beyond those provided by static information and compile-time adaptation. We consider specialization based on dynamic information like user input, architectural characteristics such as the memory hierarchy organization, and the execution profile of the application as obtained from the execution platform's performance monitoring units. One of the challenges of future execution platforms is to allow the seamless integration of these various kinds of information with information obtained from static analysis (either during ahead-of-time or just-in-time) compilation. We extend the notion of information-driven adaptation and outline the architecture of an infrastructure designed to enable information flow and adaptation through-out the life-cycle of an application. © 2011 ACM.	en
heal.journalName	ACM International Conference Proceeding Series	en
dc.identifier.doi	10.1145/2000417.2000422	en
dc.identifier.spage	46	en
dc.identifier.epage	51	en