DZone
Performance Zone
Thanks for visiting DZone today,
Edit Profile
  • Manage Email Subscriptions
  • How to Post to DZone
  • Article Submission Guidelines
Sign Out View Profile
  • Post an Article
  • Manage My Drafts
Over 2 million developers have joined DZone.
Log In / Join
  • Refcardz
  • Trend Reports
  • Webinars
  • Zones
  • |
    • Agile
    • AI
    • Big Data
    • Cloud
    • Database
    • DevOps
    • Integration
    • IoT
    • Java
    • Microservices
    • Open Source
    • Performance
    • Security
    • Web Dev
DZone > Performance Zone > Microbenchmarking Java - Compare Algorithms

Microbenchmarking Java - Compare Algorithms

Peter Karussell user avatar by
Peter Karussell
·
May. 22, 09 · Performance Zone · Interview
Like (0)
Save
Tweet
5.48K Views

Join the DZone community and get the full member experience.

Join For Free

There are a lot of microbenchmark tips out pn the internet. The intent of them differs from author to author.Today I want to show how you could compare e.g. different algorithms. You could simply do:

int COUNT = 1000000;
long firstMillis = System.currentTimeMillis();
for(int i = 0; i < COUNT; i++) {
  runAlgorithm();
}
System.out.println("Mean runtime of algorithm in seconds:"+(System.currentTimeMillis()-firstMillis) / 1000.0 / COUNT);

There are several problems with this approach, which results in very unpredictable results:

  1. Make sure the variable COUNT is high enough (if runAlgorithm() is unexpensive). Then make sure you run this code several times. Additionally you should measure the RMS either for each runAlgorithm call or at the end of this code snippet for a better comparison with other algorithms.
  2. You should turn off the JIT-compiler with specifying -Xint as a JVM option, otherwise your outcome could depend on the (unpredictable) JIT and not on your algorithm.
  3. You should start with enough memory, so specify e.g.-Xms64m -Xmx64m. Because JVM memory allocation could get time consuming.
  4. Quit all other applications or at least you should use
    long firstNanos = mx.getCurrentThreadCpuTime();
    ThreadMXBean mx = ManagementFactory.getThreadMXBean();

    instead of

    long firstMillis = System.currentTimeMillis();
  5. Avoid that the garbage collector runs while your measurement: -Xnoclassg
    But make sure you have enough memory!

After this I got relative stable results: The difference of maximal and minimal value for the result is less then 4% after 10 runs.

From http://karussell.wordpress.com

Algorithm Java (programming language)

Opinions expressed by DZone contributors are their own.

Popular on DZone

  • ETL, ELT, and Reverse ETL
  • How to Test JavaScript Code in a Browser
  • Create a Self-Service Customer Support Chatbot Without Code
  • Debugging the Java Message Service (JMS) API Using Lightrun

Comments

Performance Partner Resources

X

ABOUT US

  • About DZone
  • Send feedback
  • Careers
  • Sitemap

ADVERTISE

  • Advertise with DZone

CONTRIBUTE ON DZONE

  • Article Submission Guidelines
  • MVB Program
  • Become a Contributor
  • Visit the Writers' Zone

LEGAL

  • Terms of Service
  • Privacy Policy

CONTACT US

  • 600 Park Offices Drive
  • Suite 300
  • Durham, NC 27709
  • support@dzone.com
  • +1 (919) 678-0300

Let's be friends:

DZone.com is powered by 

AnswerHub logo