Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [cdt-core-dev] Bad scalability news

Thanks,

  I'll give it a go today and see how much luck I get.  This is the first
time
that I've actually been able to see the fruits of the Hyades project.  

Thomas

>-----Original Message-----
>From: Pierre-Alexandre Masse [mailto:pmasse@xxxxxxxxxx]
>Sent: May 17, 2004 10:09 PM
>To: cdt-core-dev@xxxxxxxxxxx
>Subject: RE: [cdt-core-dev] Bad scalability news
>
>
>Thomas,
>
>I used hyades with eclipse-2.1.0. However today, I tried a few things
>for you to see what would work. I tried the followings:
>
>1.
>Eclipse-2.1.3
>gef-2.1.3
>emf-1.1.1
>xsd-1.1.1.1
>hyades-1.3.0
>cdt-1.2
>
>2.
>Eclipse-3.0M8
>gef I20040330
>emf I20040325631
>xsd I200405131028
>hyades 3.0.0_20040331_1118
>cdt 2.0 (but there is no M8 version...)
>
>Basically both configuration worked fine. What you have to do 
>is just to
>install all these plugins. To install and configure hyades_rac (follow
>instruction in "getting_started.txt", it fairly simple and well
>explained)
>Once you install all these things, 
>- start the racserver (from hyades)
>- start you development eclipse (with a lot of memory.... e.g.:
>-Xmx256M, it depends on what you want to profile)
>- from your development environment start the workbench you want to
>profile as usual with just one additional arguments for the VM
>(-XrunpiAgent)
>- when you want to start profiling, in your development workbench go in
>your profiling perspective and in the profiling view, in the popup menu
>select attach to Java process. It should list one process... it is your
>runtime workbench even though the name is not very explicit. Follow the
>wizard, check/uncheck the different options as you wish 
>(considering the
>issue you may want to choose memory analysis...)
>- once you are done, in the profiling view toolbar, click on the green
>arrow to start collecting data, press again to stop acquiring data.
>
>You can right click in the profiling view on your process and open the
>view in which you have an interest to see the data (for memory 
>analysis,
>you can use for example the Object references view)
>
>With CDT 1.2 and hyades 1.3.0 I had an exception printed on my
>development environment workspace, but I was still able to look at the
>data...
>
>Let me know if something is not clear enough or if you have any
>question. 
>
>Pierre-Alexandre
>
>
>On Sun, 2004-05-16 at 09:58, Thomas Fletcher wrote:
>> Pierre-Alexandre,
>> 
>>   I'm curious about how you used Hyades to help with this.  I've been
>> looking
>> for a way to experiment with Hyades in a "real-world" type 
>of problem to see
>> what kind of a solution/opportunity they provide and would 
>be interested in 
>> expanding on this approach.  Also, if we can we can put it 
>to good use in
>> further
>> testing efforts on the CDT using an "all eclipse" style solution.
>> 
>> Thanks,
>>  Thomas
>> 
>> > -----Original Message-----
>> > From: Pierre-Alexandre Masse [mailto:pmasse@xxxxxxxxxx]
>> > Sent: Wednesday, May 12, 2004 12:26 PM
>> > To: cdt-core-dev@xxxxxxxxxxx
>> > Subject: Re: [cdt-core-dev] Bad scalability news
>> > 
>> > 
>> > 
>> > > Unfortunately no one here yet knows how to dump all the 
>> > objects allocated 
>> > > in the jvm. Hprof is lovely, but to nail this we really 
>> > don't want to track 
>> > > everything for all time, we just want to dump twice and see 
>> > what's new. Any 
>> > > pointers on how to do that?
>> > 
>> > I had a similar problem with DevRocket and I used hyades 
>to profile my
>> > operation. If you profile the very particular operation you are
>> > interested in (in this case opening and closing a c file I 
>> > believe) you
>> > should find the data very useful and it should point you 
>> > quickly to the
>> > non-released objects.
>> > 
>> > I hope it helps.
>> > 
>> > Pierre-Alexandre
>> > 
>> > 
>> > _______________________________________________
>> > cdt-core-dev mailing list
>> > cdt-core-dev@xxxxxxxxxxx
>> > http://dev.eclipse.org/mailman/listinfo/cdt-core-dev
>> > 
>> _______________________________________________
>> cdt-core-dev mailing list
>> cdt-core-dev@xxxxxxxxxxx
>> http://dev.eclipse.org/mailman/listinfo/cdt-core-dev
>
>_______________________________________________
>cdt-core-dev mailing list
>cdt-core-dev@xxxxxxxxxxx
>http://dev.eclipse.org/mailman/listinfo/cdt-core-dev
>


Back to the top