[MPICH2-dev] MPICH2 bugs in configure and maybe MPI_Cancel

[Philippe Combes]

Guillaume Mercier a écrit :
> Hello,
> 
> On Fri, 17 Nov 2006, Philippe Combes wrote:
> 
>>./configure --with-device=ch3:ssm --without-mpe --disable-f77
>>--disable-f90 --enable-cxx --enable-fast --enable-sharedlibs=gcc
>>--enable-strict --enable-error-checking=all --enable-error-messages=all
>>--enable-g=all --enable-debuginfo -prefix=/usr/local/mpich_dbg
>>First, after the configure runs, I still have a "#undef
>>HAVE_ERROR_CHECKING" in mpichconf.h, and MPICH_ERROR_MSG_LEVEL was set
>>to MPICH_ERROR_MSG_NONE. Strange, isn't it ?
> 
> 
> I may be wrong, but I think that --enable-fast disables error checking.
> That might explain the configure problem.

That's true ! You are right. Sorry, I have not read anything about it. I 
have checked that it disables error checking as well as error messages.

There is only that issue about shared libs left now...

>>Now let's talk about the initial errors.
>>Running my 3-node program under gdb leads to three SIGSEV. On two

>  Error encountered before initializing MPICH
> .. it says "before intializing" although these errors occur AT THE END of the execution and all communications were OK before.

It was a clue, actually. But LAM/MPI error message was more explicit: I 
have a function that calls MPI_Finalize, but it performs some MPI_Test's 
and MPI_Cancel's before. That function was called twice, and thus 
MPI_Cancel was called AFTER MPI_Finalize - silly me !

Best regards

Philippe