Introduction on performance analysis and profiling methodologies for KVM on ARM virtualization

Antonios Motakis; Alexander Spyridakis; Daniel Raho

doi:10.1117/12.2018086

28 May 2013 Introduction on performance analysis and profiling methodologies for KVM on ARM virtualization

Antonios Motakis, Alexander Spyridakis, Daniel Raho

Proceedings Volume 8764, VLSI Circuits and Systems VI; 87640N (2013) https://doi.org/10.1117/12.2018086
Event: SPIE Microtechnologies, 2013, Grenoble, France

Abstract

The introduction of hardware virtualization extensions on ARM Cortex-A15 processors has enabled the implementation of full virtualization solutions for this architecture, such as KVM on ARM. This trend motivates the need to quantify and understand the performance impact, emerged by the application of this technology. In this work we start looking into some interesting performance metrics on KVM for ARM processors, which can provide us with useful insight that may lead to potential improvements in the future. This includes measurements such as interrupt latency and guest exit cost, performed on ARM Versatile Express and Samsung Exynos 5250 hardware platforms. Furthermore, we discuss additional methodologies that can provide us with a deeper understanding in the future of the performance footprint of KVM. We identify some of the most interesting approaches in this field, and perform a tentative analysis on how these may be implemented in the KVM on ARM port. These take into consideration hardware and software based counters for profiling, and issues related to the limitations of the simulators which are often used, such as the ARM Fast Models platform.

Citation Download Citation

Antonios Motakis, Alexander Spyridakis, and Daniel Raho "Introduction on performance analysis and profiling methodologies for KVM on ARM virtualization", Proc. SPIE 8764, VLSI Circuits and Systems VI, 87640N (28 May 2013); https://doi.org/10.1117/12.2018086

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available