Bug 493 - performance regressed between 2.10.x and 2.6.x
Summary: performance regressed between 2.10.x and 2.6.x
Alias: None
Product: Runtime
Classification: Mono
Component: JIT ()
Version: unspecified
Hardware: All All
: --- normal
Target Milestone: ---
Assignee: Bugzilla
Depends on:
Reported: 2011-08-27 18:54 UTC by Jonathan Shore
Modified: 2011-08-29 07:42 UTC (History)
3 users (show)

Is this bug a regression?: ---
Last known good build:

The struct and class creation benchmark (2.29 KB, application/octet-stream)
2011-08-27 18:54 UTC, Jonathan Shore

Notice (2018-05-24): bugzilla.xamarin.com is now in read-only mode.

Please join us on Visual Studio Developer Community and in the Xamarin and Mono organizations on GitHub to continue tracking issues. Bugzilla will remain available for reference in read-only mode. We will continue to work on open Bugzilla bugs, copy them to the new locations as needed for follow-up, and add the new items under Related Links.

Our sincere thanks to everyone who has contributed on this bug tracker over the years. Thanks also for your understanding as we make these adjustments and improvements for the future.

Please create a new report on GitHub or Developer Community with your current version information, steps to reproduce, and relevant error messages or log files if you are hitting an issue that looks similar to this resolved bug and you do not yet see a matching new report.

Related Links:

Description Jonathan Shore 2011-08-27 18:54:53 UTC
Created attachment 199 [details]
The struct and class creation benchmark

Description of Problem:

A simple benchmark testing the use of struct and class objects at high frequency is 20% - 50% slower in 2.10.x vs 2.6.x.  Struct creation has become more expensive in 2.10.x.

Steps to reproduce the problem:
1. Run enclosed simple benchmark on 2.6.x repeatedly
2. Run enclosed simple benchmark on 2.10.x repeatedly

Actual Results:

benchmark 20 - 50% slower on 2.10.x, depending on platform.

Expected Results:

benchmark the same or better on 2.10.x

How often does this happen? 


Additional Information:

See enclosed test program
Comment 1 Zoltan Varga 2011-08-28 12:02:54 UTC
For me, the 'struct' case is exactly the same for 2.6 and 2.10/HEAD on linux, the 'class' case is a bit slower, probably due to some GC changes.
Comment 2 Rodrigo Kumpera 2011-08-29 07:42:45 UTC
Same here, struct performance is the same.

Two things to note, boehm perf on OSX regressed between 2.6 and 2.10 since we disabled parallel marking as it's buggy.

OSX boehm allocation performance is pretty bad since it won't use managed allocators. This limitation won't be changed since all our efforts are not on sgen.