Name
Abstract | ||
|---|---|---|
Understanding the performance of a multi-threaded application is difficult. The threads interfere when they access the same shared resource, which slows down their execution. Unfortunately, current profiling tools report the hardware components or the synchronization primitives that saturate, but they cannot tell if the saturation is the cause of a performance bottleneck. In this paper, we propose a holistic metric able to pinpoint the blocks of code that suffer interference the most, regardless of the interference cause. Our metric uses performance variation as a universal indicator of interference problems. With an evaluation of 27 applications we show that our metric can identify interference problems caused by six different kinds of interference in nine applications. We are able to easily remove seven of the bottlenecks, which leads to a performance improvement of up to nine times. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1109/TPDS.2019.2927481 | IEEE Transactions on Parallel and Distributed Systems |
Keywords | Field | DocType |
Real-time systems,Energy storage,Renewable energy sources,Power system stability,Supply and demand,Generators,Production | Bottleneck,Multithreading,Synchronization,Computer science,Profiling (computer programming),Thread (computing),Interference (wave propagation),Shared resource,Distributed computing,Performance improvement | Journal |
Volume | Issue | ISSN |
30 | 12 | 1045-9219 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
8 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Mohamed Said Mosli Bouksiaa | 1 | 0 | 0.34 |
| François Trahay | 2 | 33 | 7.51 |
| Alexis Lescouet | 3 | 0 | 0.34 |
| Gauthier Voron | 4 | 0 | 0.34 |
| Rémi Dulong | 5 | 0 | 0.34 |
| Amina Guermouche | 6 | 176 | 8.73 |
| Elisabeth Brunet | 7 | 1 | 1.38 |
| Gaël Thomas | 8 | 256 | 16.95 |