Skip to content
This repository

JMeter performance comparison

This page used to display a comparison based on JMeter's benchmark described there:

Accidentally, this comparison was flawed (Gatling's HTTP Caching feature is enabled by default but not JMeter's one), and it was only reported one year after it was published. One more thing was also that this page was not maintained, and that both tools progressed a lot in the mean time.

We wanted to use a shared test case, so that it would be easy to reproduce and maybe report bugs. It's just a pity it was only reported so late.

Performance of both tool are currently (as of Gatling 1.5.3 and JMeter 2.11) roughly the same for this test case:

  • hammering a local tomcat (number of connections is too high for Tomcat, so that it replies with Connection: Close), so basically the test is actually testing how fast the underlying HTTP engine can open new connections.
  • not performing any computation/validation on the response, so the stress tool doesn't actually do much

At this point, one can only conclude that under those conditions, both tools have eliminated much of their overhead on top of their respective underlying HTTP engine.

Until we decide (and find time) so build a feature rich and reproductible benchmark, this page will stay as is and will no longer be linked.

Something went wrong with that request. Please try again.