Loading of rb files is much slower in JRuby than in MRI (2x-5x)
---------------------------------------------------------------

                 Key: JRUBY-4263
                 URL: http://jira.codehaus.org/browse/JRUBY-4263
             Project: JRuby
          Issue Type: Bug
          Components: Parser, Performance
         Environment: latest jruby from master branch
            Reporter: Vladimir Sizikov
            Assignee: Thomas E Enebo


I've just comitted the bench/bench_load.rb benchmark, and we are 2x-5x slower 
then MRI (varies a bit between platforms). This is one of those issues 
identified while trying to make JRuby startup faster.

Essentially, cold load/require is pretty bad in JRuby. Hopefully, this 
benchmark could help out to narrow some cases.
When just loading/requiring 'fileutils' takes 200ms, that's bad.

{noformat}
#>jruby bench/bench_load.rb
                                         user     system      total        real
  1 load 'fileutils-like'            0.186000   0.000000   0.186000 (  0.157000)
 1K load 'fileutils-like'           12.336000   0.000000  12.336000 ( 12.336000)
 1K load 'rational'                  3.698000   0.000000   3.698000 (  3.698000)
10K require 'non-existing'           5.024000   0.000000   5.024000 (  5.024000)

#>ruby bench/bench_load.rb
                                         user     system      total        real
  1 load 'fileutils-like'            0.000000   0.020000   0.020000 (  0.017378)
 1K load 'fileutils-like'            1.870000   1.790000   3.660000 (  3.662594)
 1K load 'rational'                  0.710000   0.700000   1.410000 (  1.461065)
10K require 'non-existing'           0.280000   0.610000   0.890000 (  0.889232)

{noformat}

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: 
http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Reply via email to