Skip to Content.
Sympa Menu

charm - Re: [charm] backtraces

charm AT lists.cs.illinois.edu

Subject: Charm++ parallel programming system

List archive

Re: [charm] backtraces


Chronological Thread 
  • From: Jonathan Lifflander <jliffl2 AT illinois.edu>
  • To: Nicolas Bock <nicolasbock AT gmail.com>
  • Cc: "charm AT cs.uiuc.edu" <charm AT cs.uiuc.edu>
  • Subject: Re: [charm] backtraces
  • Date: Mon, 29 Jul 2013 15:42:34 -0500
  • List-archive: <http://lists.cs.uiuc.edu/pipermail/charm/>
  • List-id: CHARM parallel programming system <charm.cs.uiuc.edu>

A couple things that I should mention:

- CharmDebug is a grad project that still needs some polish...so it
might not be really easy to use out of the box.
- You need to use the CCS client/server interface with it, which
connects to a port that you specify, etc. This should be in a manual
(I need to dig a little to find it)
- It currently does not support SMP mode.

On Mon, Jul 29, 2013 at 3:27 PM, Nicolas Bock
<nicolasbock AT gmail.com>
wrote:
> I must be doing something completely wrong, but charmdebug goes up in flames
> with an exception:
>
> ServThread terminated
> Finished running parallel program
> java.net.ConnectException: Connection refused
> at sun.nio.ch.Net.connect0(Native Method)
> at sun.nio.ch.Net.connect(Net.java:465)
> at sun.nio.ch.Net.connect(Net.java:457)
> at sun.nio.ch.SocketChannelImpl.connect(SocketChannelImpl.java:639)
>
> And I can run charmrun on the executable fine, either with rsh or ssh.
>
> Thanks,
>
> nick
>
>
>
> On Mon, Jul 29, 2013 at 1:55 PM, Jonathan Lifflander
> <jliffl2 AT illinois.edu>
> wrote:
>>
>> Normal GDB backtraces are not very useful when the system is
>> delivering a message and fails. To ameliorate this, we have developed
>> a tool called CharmDebug which allows you to inspect the queue on each
>> processor, change the order of delivery, etc.
>>
>> Here is the link to the research area on our website, where you can
>> find many papers:
>>
>> https://charm.cs.illinois.edu/research/parallel_debug
>>
>> The software is available to download on the website under the software
>> section.
>>
>> I'm not exactly sure how a missing extern (which should be caught at
>> link time), ended up producing a runtime error. Could you provide a
>> little more information?
>>
>> Thanks,
>>
>> Jonathan
>>
>> On Mon, Jul 29, 2013 at 2:41 PM, Nicolas Bock
>> <nicolasbock AT gmail.com>
>> wrote:
>> > It turns out that this particular problem was caused by a missing
>> > "extern
>> > module ...;" statement in one of the interface files. As a feature
>> > request:
>> > it would be great if the compiler caught omissions like these.
>> >
>> > Thanks,
>> >
>> > nick
>> >
>> >
>> > On Mon, Jul 29, 2013 at 12:13 PM, Nicolas Bock
>> > <nicolasbock AT gmail.com>
>> > wrote:
>> >>
>> >> Hi,
>> >>
>> >> I keep getting segfaults deep inside the RTS. Unfortunately I haven't
>> >> found a good way yet to connect the backtraces to my code, since the
>> >> callstack doesn't actually involve anything from my code. Is there a
>> >> recommended technique to debug these kinds of situations?
>> >>
>> >> Thanks,
>> >>
>> >> nick
>> >>
>> >
>> >
>> > _______________________________________________
>> > charm mailing list
>> > charm AT cs.uiuc.edu
>> > http://lists.cs.uiuc.edu/mailman/listinfo/charm
>> >
>
>




Archive powered by MHonArc 2.6.16.

Top of Page