MyUT: Design and implementation of efficient user-level thread management for improving cache utilization

Authors:
Inhyuk Kim;Eunhwan Shin;Junghan Kim;Young Ik Eom
Affiliations:
School of Information and Communication Eng., Sungkyunkwan University, Suwon, Gyeonggi-do, Korea;School of Information and Communication Eng., Sungkyunkwan University, Suwon, Gyeonggi-do, Korea;School of Information and Communication Eng., Sungkyunkwan University, Suwon, Gyeonggi-do, Korea;School of Information and Communication Eng., Sungkyunkwan University, Suwon, Gyeonggi-do, Korea
Venue:
ICCSA'11 Proceedings of the 2011 international conference on Computational science and Its applications - Volume Part V
Year:
2011

Citing 10
Cited 0

Cache performance of operating system and multiprogramming workloads

ACM Transactions on Computer Systems (TOCS)
The effect of context switches on cache performance

ASPLOS IV Proceedings of the fourth international conference on Architectural support for programming languages and operating systems
Scheduler activations: effective kernel support for the user-level management of parallelism

SOSP '91 Proceedings of the thirteenth ACM symposium on Operating systems principles
Exploiting choice: instruction fetch and issue on an implementable simultaneous multithreading processor

ISCA '96 Proceedings of the 23rd annual international symposium on Computer architecture
Quantifying the cost of context switch

Proceedings of the 2007 workshop on Experimental computer science
A scheduling framework for general-purpose parallel languages

Proceedings of the 13th ACM SIGPLAN international conference on Functional programming
Towards practical page coloring-based multicore cache management

Proceedings of the 4th ACM European conference on Computer systems
Composing parallel software efficiently with lithe

PLDI '10 Proceedings of the 2010 ACM SIGPLAN conference on Programming language design and implementation
Tessellation: space-time partitioning in a manycore client OS

HotPar'09 Proceedings of the First USENIX conference on Hot topics in parallelism
FlexSC: flexible system call scheduling with exception-less system calls

OSDI'10 Proceedings of the 9th USENIX conference on Operating systems design and implementation

Quantified Score

Hi-index	0.00

Visualization

Abstract

The appearance of the multicore processors and the advancement of multithread programming have lead the new paradigm of the system optimization. Especially, the cache performance has been known as the one of the most important factor of the system optimization. The user-level thread management, the unvirtualized thread and the exception-less system call are introduced to improve the cache utilization of the multithread programming and parallel programming. However, these schemes have some limitations on applications domain. So, we propose the general purpose user-level thread management scheme to reduce the context-switch, CPU-migration and synchronous system call which pollute the amount of caches. We show evaluation of our system on the three workloads. We show the performance improvements of about 10-20% in respect of the CPU, memory and IO intensive workloads and analyze the effects of the three policies and techniques through the experiments.