logo
CanyamAI summaries for academic research
language
MobileMobile
  • Phone FeatureWeChat Mini Program
EnglishEnglish
noticeNotice
BlogBlog
delete
arrow
Return

Self-tuning systems

delete1999-01-01
delete15
PRE
AI
D
Dror G. Feitelson
M
Mor Naaman
DOI：10.1109/52.754053delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
Modern operating systems are highly parameterized, allowing system administrators to tune them to optimize the local workload. However, this is difficult and time-consuming. The authors propose a mechanism to automate this process by running simulations of system performance for Various para meter values in place of the system's idle loop. The simulations are driven by log files containing information about the local workload, and genetic algorithms are used to search for the optimal parameter values. The parameter values they found reduced fragmentation and salvaged a quarter of the computing cycles that were lost when using the defaults.

Journal

IEEE Software cover
IEEE Software
IF:
3
Papers:
3.1K
Citations:
3.6K

Organization

No organization information available