Achieved IPC performance (still the foundation for extensibility)
Authors
IBM Watson Research Center
The University of New South Wales
Sydney 2052
Australia
Abstract
Extensibility can be based on cross-address-space communication or on grafting application-specific modules into the operating system. For comparing both approaches, we need to explore the best achievable performance in both models. This paper reports the achieved performance of cross-address-space communication for the L4 \mu-kernel on Intel Pentium, Mips R4600 and DEC Alpha. The direct costs range from 45 cycles (alpha) to 121 cycles (Pentium). Since only 2.3% of the L1 cache are required (Pentium), the average indirect costs are not to be expected much higher.
BibTeX Entry
@inproceedings{Liedtke_ESHHIJ_97, author = {Jochen Liedtke and Kevin Elphinstone and Sebastian Sch{\"o}nberg and Herrman H{\"a}rtig and Gernot Heiser and Nayeem Islam and Trent Jaeger}, month = may, year = {1997}, title = {Achieved {IPC} Performance (Still the Foundation for Extensibility)}, address = {Cape Cod, MA, USA}, pages = {28--31}, booktitle = {Proceedings of the 6th Workshop on Hot Topics in Operating Systems (HotOS)}, paperurl = {https://ts.data61.csiro.au/publications/papers/Liedtke_ESHHIJ_97.pdf} }