CPL - Chalmers Publication Library
| Utbildning | Forskning | Styrkeområden | Om Chalmers | In English In English Ej inloggad.

Comparing scalability prediction strategies on an SMP of CMPs

K. Singh ; M. Curtis-Maury ; Sally A McKee (Institutionen för data- och informationsteknik, Datorteknik (Chalmers)) ; F. Blagojevic ; D.S. Nikolopoulos ; B.R. De Supinski ; M. Schulz
Lecture Notes in Computer Science. 16th International Euro-Par Conference on Parallel Processing, Euro-Par 2010, Ischia, 31 August-3 September 2010 (0302-9743). Vol. 6271 (2010), p. 143-155.
[Konferensbidrag, refereegranskat]

Diminishing performance returns and increasing power consumption of single-threaded processors have made chip multiprocessors (CMPs) an industry imperative. Unfortunately, poor software/hardware interaction and bottlenecks in shared hardware structures can prevent scaling to many cores. In fact, adding a core may harm performance and increase power consumption. Given these observations, we compare two approaches to predicting parallel application scalability: multiple linear regression and artificial neural networks (ANNs). We throttle concurrency to levels with higher predicted power/performance efficiency. We perform experiments on a state-of-the-art, dual-processor, quad-core platform, showing that both methodologies achieve high accuracy and identify energy-efficient concurrency levels in multithreaded scientific applications. The ANN approach has advantages, but the simpler regression-based model achieves slightly higher accuracy and performance. The approaches exhibit median error of 7.5% and 5.6%, and improve performance by an average of 7.4% and 9.5%, respectively.



Denna post skapades 2012-02-10. Senast ändrad 2016-03-22.
CPL Pubid: 155017

 

Läs direkt!


Länk till annan sajt (kan kräva inloggning)


Institutioner (Chalmers)

Institutionen för data- och informationsteknik, Datorteknik (Chalmers)

Ämnesområden

Information Technology

Chalmers infrastruktur