Bug 65187 - [tests] Negative differences between cpu live results of a method with wait/sleep
[tests] Negative differences between cpu live results of a method with wait/s...
Status: NEW
Product: profiler
Classification: Unclassified
Component: Base
5.x
All All
: P3 (vote)
: 7.2
Assigned To: Tomas Hurka
issues@profiler
: VISUALVM
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2005-09-27 13:27 UTC by ehucka
Modified: 2011-11-02 20:22 UTC (History)
0 users

See Also:
Issue Type: DEFECT
:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description ehucka 2005-09-27 13:27:39 UTC
Live results of cpu profiling (mostly of code with wait/sleep) are nonlogicaly
changing. 
I assume that self time of a method always increasing but there are sometimes
values less than a step before or sometimes it jump to zero. After the
application finished it shows the right value.

example: jfluid/test/functional/data/projects/j2se-simple, class: WaitingTest,
method: method1000()
Comment 1 iformanek 2005-10-21 11:04:38 UTC
This has been fixed by bugfix for issue 67197
Comment 2 ehucka 2005-11-01 12:14:27 UTC
I can reproduce it with code using wait/sleep manualy and by automated tests in
dev build 051101.
Comment 3 iformanek 2005-11-02 14:09:22 UTC
Fixed in M10.
This was a sideeffect of removing cleansing of time at the get results moment.
Comment 4 ehucka 2005-11-08 09:22:18 UTC
Some automated tests are still failing but the 'fluctuations' are only in nano
seconds.
So I can rewrite automated tests to ignore nano seconds but it is strange that
there can occure such case.
Comment 5 J Bachorik 2007-01-09 07:34:05 UTC
is this issue still valid for profiler trunk?
Comment 6 ehucka 2007-01-09 08:09:56 UTC
Automated tests on solaris sparc last week failed by this issue.
It measured:
method               invocations time(us)
1. step
simple.CPU.test500() 3           1525180
2. step
simple.CPU.test500() 4           1525179

where 'step' is update of live results.
Comment 7 Tomas Hurka 2007-01-09 08:12:37 UTC
Ok, I will investgate.
Comment 8 Tomas Hurka 2007-11-07 10:36:39 UTC
Changing target milestone to dev, since NetBeans 6.0 is in high resistance mode.


By use of this website, you agree to the NetBeans Policies and Terms of Use. © 2012, Oracle Corporation and/or its affiliates. Sponsored by Oracle logo