From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gdb-return-23247-listarch-gdb=sources.redhat.com@sourceware.org>
Received: (qmail 2881 invoked by alias); 16 Nov 2005 09:19:55 -0000
Received: (qmail 2873 invoked by uid 22791); 16 Nov 2005 09:19:51 -0000
Received: from nproxy.gmail.com (HELO nproxy.gmail.com) (64.233.182.200)
    by sourceware.org (qpsmtpd/0.30-dev) with ESMTP; Wed, 16 Nov 2005 09:19:51 +0000
Received: by nproxy.gmail.com with SMTP id l36so387655nfa
        for <gdb@sources.redhat.com>; Wed, 16 Nov 2005 01:19:48 -0800 (PST)
Received: by 10.48.249.12 with SMTP id w12mr300698nfh;
        Wed, 16 Nov 2005 01:19:48 -0800 (PST)
Received: by 10.48.236.16 with HTTP; Wed, 16 Nov 2005 01:19:48 -0800 (PST)
Message-ID: <159126900511160119o24697a0an@mail.gmail.com>
Date: Wed, 16 Nov 2005 09:19:00 -0000
From: Konstantin Karganov <karganov@gmail.com>
To: Vladimir Prus <ghost@cs.msu.su>
Subject: Re: libGDB and gdbserver questions
Cc: gdb@sources.redhat.com
In-Reply-To: <dlcu2c$ct1$1@sea.gmane.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
References: <bba22c370511150448t1ab0102fi4f4e1e38979a512e@mail.gmail.com>
	 <dlcu2c$ct1$1@sea.gmane.org>
Mailing-List: contact gdb-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:gdb-subscribe@sourceware.org>
List-Archive: <http://sourceware.org/ml/gdb/>
List-Post: <mailto:gdb@sourceware.org>
List-Help: <mailto:gdb-help@sourceware.org>, <http://sourceware.org/ml/#faqs>
Sender: gdb-owner@sourceware.org
X-SW-Source: 2005-11/txt/msg00317.txt.bz2

> Then each copy of gdb would connect to a "redirector" you can write, that
> will basically forward all packets to invididual instances of gdbserver.
MPICH already has such wrapper (I was sure it was, but couldn't find
it yesterday...):
It has a "parallel gdb" version called mpigdb:
http://www-unix.mcs.anl.gov/mpi/mpich/docs/userguide/node26.htm#Node29

It wraps the stdio and allows to send gdb commands either to all nodes
or to one specified node. The worst thing is that it works
synchronously - when you issue "step" to one node you can not issue
other commands (even switching to another node) until the step
completes.

As for me, such solution is not suitable even for small MPI programs,
because of process interaction and blocking of the individual steps on
MPI_Recv() and similar.
Anyhow, I don't know the interactive debugger that is capable of
debugging 1000 and more processes.

--
Best regards,
Konstantin.