DMCA
Thread Tailor: Dynamically Weaving Threads Together for Efficient, Adaptive Parallel Applications
| Citations: | 16 - 2 self |
BibTeX
@MISC{Lee_threadtailor:,
author = {Janghaeng Lee and Haicheng Wu and Madhumitha Ravich and Nathan Clark},
title = {Thread Tailor: Dynamically Weaving Threads Together for Efficient, Adaptive Parallel Applications},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Extracting performance from modern parallel architectures requires that applications be divided into many different threads of execution. Unfortunately selecting the appropriate number of threads for an application is a daunting task. Having too many threads can quickly saturate shared resources, such as cache capacity or memory bandwidth, thus degrading performance. On the other hand, having too few threads makes inefficient use of the resources available. Beyond static resource assignment, the program inputs and dynamic system state (e.g., what other applications are executing in the system) can have a significant impact on the right number of threads to use for a particular application. To address this problem we present the Thread Tailor, a dynamic system that automatically adjusts the number of threads in an application to optimize system efficiency. The Thread Tailor leverages offline analysis to estimate what type of threads will exist at runtime and the communication patterns between them. Using this information Thread Tailor dynamically combines threads to better suit the needs of the target system. Thread Tailor adjusts not only to the architecture, but also other applications in the system, and this paper demonstrates that this type of adjustment can lead to significantly better use of thread-level parallelism in real-world architectures.
Keyphrases
thread tailor thread together adaptive parallel application inefficient use right number real-world architecture dynamic system thread-level parallelism many thread cache capacity beyond static resource assignment dynamic system state significant impact offline analysis communication pattern target system shared resource information thread tailor appropriate number many different thread program input modern parallel architecture memory bandwidth system efficiency particular application
