COMPROF and COMPLACE: shared-memory communication profiling and automated thread placement via dynamic binary instrumentation

Ryan Kirkpatrick*, Christopher Mark Brown, Vladimir Janjic

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents COMPROF and COMPLACE, a novel profiling tool and thread placement technique for shared-memory architectures that requires no recompilation or user intervention. We use dynamic binary instrumentation to intercept memory operations and estimate inter-thread communication overhead, deriving (and possibly visualising) a communication graph of data-sharing between threads. We then use this graph to map threads to cores in order to optimise memory traffic through the memory system.
Different paths through a system's memory hierarchy have different latency, throughput and energy properties, COMPLACE exploits this heterogeneity to provide automatic performance and energy improvements for multi-threaded programs. We demonstrate COMPLACE on the NAS Parallel Benchmark (NPB) suite where, using our technique, we are able to achieve improvements of up to 12% in the execution time and up to 10% in the energy consumption (compared to default Linux scheduling) while not requiring any modification or recompilation of the application code.
Original languageEnglish
Title of host publication29th IEEE International Conference on High Performance Computing, Data, and Analytics
DOIs
Publication statusAccepted/In press - 15 Nov 2022
Event29th IEEE International Conference on High Performance Computing, Data, and Analytics (HIPC) - Bangalore, India
Duration: 18 Dec 202221 Dec 2022
Conference number: 29
https://hipc.org/

Publication series

NameIEEE International Conference on High Performance Computing, Data, and Analytics
ISSN (Print)1094-7256
ISSN (Electronic)2640-0316

Conference

Conference29th IEEE International Conference on High Performance Computing, Data, and Analytics (HIPC)
Abbreviated titleHIPC
Country/TerritoryIndia
CityBangalore
Period18/12/2221/12/22
Internet address

Keywords

  • NUMA
  • Thread Placement
  • Data Placement
  • Cache Optimisation
  • Energy Optimization
  • Refactoring

Fingerprint

Dive into the research topics of 'COMPROF and COMPLACE: shared-memory communication profiling and automated thread placement via dynamic binary instrumentation'. Together they form a unique fingerprint.

Cite this