Quantifying the effects of scheduling on IPC performance
Authors
NICTA, Sydney, Australia
UNSW, Australia
Abstract
This thesis investigates three scheduling performance optimisations used in the L4 kernel and the trades-offs associated with them. The three performance optimisations: “lazy-queueing”, “direct process switching” and “FIFO queue ordering” each improve best-effort throughput of the scheduler, but decrease real-time performance of the kernel. We attempt to quantify both the advantages and the disadvantages of each.
BibTeX Entry
@mastersthesis{Greenaway:bsc, month = jun, school = {School of Computer Science and Engineering}, author = {David Greenaway}, title = {Quantifying the Effects of Scheduling on {IPC} Performance}, address = {Sydney, Australia}, year = {2007} }