Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [tracecompass-dev] Strange behavior with flame graph
Apologies - I thought I had tried that and didn't get the update, but I had failed to update the Target Platform, so I did not see the update. When I updated the Target Platform, the update came in.

After some testing, it looks like there might still be one outstanding issue: when I switch back and forth between the "Per Thread" and "Aggregate" views on the Flame Graph, I notice that the functions at the root level show an increasing number of calls statistic.  The functions below the root level seem to be correct.  So, I guess a question is whether the same fix needs to be applied at the root level.

Let me know if you need more details, but maybe Bernd will see the issue right away.

Rocky

On Thu, Oct 12, 2017 at 11:28 AM, Matthew Khouzam <matthew.khouzam@xxxxxxxxxxxx> wrote:
How about using the stable update site:

http://download.eclipse.org/tracecompass/stable/repository/

If you are using eclipse, add this to "Install new Software".

If you are using an RCP, add this

http://download.eclipse.org/tracecompass/stable/rcp-repository/

That being said, the latest RCP should have "auto notify" updates
enabled. Did you not get an update notification?

You can also get the latest RCP here.
http://download.eclipse.org/tracecompass/stable/rcp/

This is the executable with later patches.

Please tell us if it works.


Matthew.



On 17-10-12 01:06 PM, Rocky Dunlap - NOAA Affiliate wrote:
> I guess what I am asking for is a 3.1.1 release so that clients can
> pull down the update.
>
> Rocky
>
> On Thu, Oct 12, 2017 at 11:01 AM, Matthew Khouzam
> <matthew.khouzam@xxxxxxxxxxxx <mailto:matthew.khouzam@ericsson.com>>
> wrote:
>
>     Bernd is way ahead of us, https://git.eclipse.org/r/#/c/106227/
>     <https://git.eclipse.org/r/#/c/106227/>
>
>     Stable 3.1 has the patch in it. However, I don't know if there
>     will be a
>     Trace Compass 3.1.1 for workgroups.
>
>     Here are some binaries with the fix.
>
>     https://hudson.eclipse.org/tracecompass/view/All/job/tracecompass-stable-nightly/lastSuccessfulBuild/artifact/
>     <https://hudson.eclipse.org/tracecompass/view/All/job/tracecompass-stable-nightly/lastSuccessfulBuild/artifact/>
>
>
>     On 17-10-12 12:40 PM, Rocky Dunlap - NOAA Affiliate wrote:
>     > Geneviève,
>     >
>     > Thanks very much for committing a fix for this so quickly.
>     > Would it be possible to push out a patch of 3.1 including this
>     change?
>     >
>     > Rocky
>     >
>     > On Thu, Sep 28, 2017 at 12:48 PM, Genevieve Bastien
>     > <gbastien+lttng@xxxxxxxxxxxx <mailto:gbastien%2Blttng@versatic.net>
>     <mailto:gbastien+lttng@versatic.net
>     <mailto:gbastien%2Blttng@versatic.net>>> wrote:
>     >
>     >     Hi Rocky,
>     >
>     >
>     >     This bug is indeed for the same reason as the one we had in the
>     >     incubator: grouping merges the callees, but they are being
>     merged
>     >     into the object from the first element! It should make a copy
>     >     first. Too bad it made it to 3.1... :(
>     >
>     >
>     >     Let's see how fast we can fix it.
>     >
>     >
>     >     Geneviève
>     >
>     >
>     >
>     >     On 2017-09-28 01:27 PM, Rocky Dunlap - NOAA Affiliate wrote:
>     >>     Thanks Matthew.
>     >>
>     >>     Here's the bug:
>     >>     https://bugs.eclipse.org/bugs/show_bug.cgi?id=525334
>     <https://bugs.eclipse.org/bugs/show_bug.cgi?id=525334>
>     >>     <https://bugs.eclipse.org/bugs/show_bug.cgi?id=525334
>     <https://bugs.eclipse.org/bugs/show_bug.cgi?id=525334>>
>     >>
>     >>     It would certainly help out to backport the current fix in
>     >>     incubator to master.
>     >>
>     >>     Rocky
>     >>
>     >>     On Thu, Sep 28, 2017 at 11:12 AM, Matthew Khouzam
>     >>     <matthew.khouzam@xxxxxxxxxxxx
>     <mailto:matthew.khouzam@ericsson.com>
>     >>     <mailto:matthew.khouzam@ericsson.com
>     <mailto:matthew.khouzam@ericsson.com>>> wrote:
>     >>
>     >>         Wow, that's really cool!
>     >>
>     >>         This is pure conjecture, I think we may have a problem with
>     >>         adding
>     >>         intervals, we are adding durations but they include t0
>     so it may
>     >>         increment by one each element.
>     >>
>     >>         Would you be able to open a bug with this? (Also if you can
>     >>         contribute a
>     >>         patch to fix it, we can review it quickly. :))
>     >>
>     >>         For the third item, I saw that in the incubator and
>     Genevieve
>     >>         fixed it,
>     >>         maybe we could backport that fix to master. I think
>     it's this.
>     >>         https://git.eclipse.org/r/#/c/102326/
>     <https://git.eclipse.org/r/#/c/102326/>
>     >>         <https://git.eclipse.org/r/#/c/102326/
>     <https://git.eclipse.org/r/#/c/102326/>>
>     >>
>     >>         Thanks.
>     >>
>     >>         Matthew
>     >>
>     >>         On 17-09-28 01:00 PM, Rocky Dunlap - NOAA Affiliate wrote:
>     >>         > We have created a custom trace that extends
>     CtfTmfTrace and an
>     >>         > analysis by extending CallStackAnalysis and
>     >>         CallStackStateProvider.
>     >>         > We are using TraceCompass 3.1.   The analysis appears
>     to be
>     >>         working
>     >>         > correctly and the call stack view looks good.  I also
>     >>         noticed that we
>     >>         > get the Flame Graph "for free" since we have extended
>     >>         > CallStackAnalysis.  However, we are seeing some strange
>     >>         behavior with
>     >>         > the Flame Graph.
>     >>         >
>     >>         > When I first load the trace, the "per thread" Flame
>     Graph looks
>     >>         > correct.  See below where there are four threads
>     (each in a
>     >>         separate
>     >>         > process) and notice that the function named "[GCS]
>     run #1"
>     >>         has 192
>     >>         > calls in each thread.
>     >>         >
>     >>         > Inline image 1
>     >>         >
>     >>         > Then, I immediately switch to "aggregate threads
>     view" and
>     >>         see that
>     >>         > the total number of calls for "[GCS] run #1" is 772,
>     but I
>     >>         think this
>     >>         > should be 192*4 = 768.
>     >>         >
>     >>         > Inline image 2
>     >>         >
>     >>         > Then, the really strange part is when I switch back
>     to "per
>     >>         thread"
>     >>         > view I see that one of the threads shows 1355 calls to
>     >>         "[GCS] run
>     >>         > #1".  I would expect this view to be identical to the
>     first
>     >>         time
>     >>         > viewing the "per thread" view.
>     >>         >
>     >>         > Inline image 3
>     >>         >
>     >>         > Now, when I switch back to the aggregate view, the total
>     >>         number of
>     >>         > calls for this function has gone up to 2530.  I am
>     >>         suspicious that
>     >>         > this is not the expected behavior.  However, since I have
>     >>         created a
>     >>         > custom CtfTmfTrace, maybe I do not have something set
>     up in
>     >>         the state
>     >>         > system / attribute tree.  The strange this is that
>     the call
>     >>         stack view
>     >>         > itself works great.  Keep in mind that these four threads
>     >>         are in four
>     >>         > different processes--no idea if that may shed light
>     on the
>     >>         strange
>     >>         > behavior.
>     >>         >
>     >>         > Any ideas?
>     >>         >
>     >>         > Rocky
>     >>         >
>     >>         >
>     >>         > _______________________________________________
>     >>         > tracecompass-dev mailing list
>     >>         > tracecompass-dev@xxxxxxxxxxx
>     <mailto:tracecompass-dev@eclipse.org>
>     >>         <mailto:tracecompass-dev@eclipse.org
>     <mailto:tracecompass-dev@eclipse.org>>
>     >>         > To change your delivery options, retrieve your password, or
>     >>         unsubscribe from this list, visit
>     >>         >
>     https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>
>     >>       
>      <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>>
>     >>
>     >>         _______________________________________________
>     >>         tracecompass-dev mailing list
>     >>         tracecompass-dev@xxxxxxxxxxx
>     <mailto:tracecompass-dev@eclipse.org>
>     >>         <mailto:tracecompass-dev@eclipse.org
>     <mailto:tracecompass-dev@eclipse.org>>
>     >>         To change your delivery options, retrieve your password, or
>     >>         unsubscribe from this list, visit
>     >>       
>      https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>
>     >>       
>      <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>>
>     >>
>     >>
>     >>
>     >>
>     >>     _______________________________________________
>     >>     tracecompass-dev mailing list
>     >>     tracecompass-dev@xxxxxxxxxxx
>     <mailto:tracecompass-dev@eclipse.org>
>     <mailto:tracecompass-dev@eclipse.org
>     <mailto:tracecompass-dev@eclipse.org>>
>     >>     To change your delivery options, retrieve your password, or unsubscribe from this list, visit
>     >>     https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>
>     >>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>>
>     >
>     >
>     >     _______________________________________________
>     >     tracecompass-dev mailing list
>     >     tracecompass-dev@xxxxxxxxxxx
>     <mailto:tracecompass-dev@eclipse.org>
>     <mailto:tracecompass-dev@eclipse.org
>     <mailto:tracecompass-dev@eclipse.org>>
>     >     To change your delivery options, retrieve your password, or
>     >     unsubscribe from this list, visit
>     >     https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>
>     >     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>>
>     >
>     >
>     >
>     >
>     > _______________________________________________
>     > tracecompass-dev mailing list
>     > tracecompass-dev@xxxxxxxxxxx <mailto:tracecompass-dev@eclipse.org>
>     > To change your delivery options, retrieve your password, or
>     unsubscribe from this list, visit
>     > https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>
>
>     _______________________________________________
>     tracecompass-dev mailing list
>     tracecompass-dev@xxxxxxxxxxx <mailto:tracecompass-dev@eclipse.org>
>     To change your delivery options, retrieve your password, or
>     unsubscribe from this list, visit
>     https://dev.eclipse.org/mailman/listinfo/tracecompass-dev
>     <https://dev.eclipse.org/mailman/listinfo/tracecompass-dev>
>
>
>
>
> _______________________________________________
> tracecompass-dev mailing list
> tracecompass-dev@xxxxxxxxxxx
> To change your delivery options, retrieve your password, or unsubscribe from this list, visit
> https://dev.eclipse.org/mailman/listinfo/tracecompass-dev

_______________________________________________
tracecompass-dev mailing list
tracecompass-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/tracecompass-dev

Back to the top