LAM/MPI logo

LAM/MPI General User's Mailing List Archives

  |   Home   |   Download   |   Documentation   |   FAQ   |   all just in this list

From: Jeff Whitaker (jswhit_at_[hidden])
Date: 2006-06-26 10:44:01


Brian Barrett wrote:
> On Jun 19, 2006, at 10:31 AM, Jeff Whitaker wrote:
>
>> Brian: Thanks for the explanation. It's unfortunate, since MPI_SUM
>> with real8 is a fairly common construct, at least in the codes I deal
>> with. I've tried openmpi, but am encountering mysterious crashes
>> on OS
>> X. mpich2 seems to work fine though, so I'll go with that for now.
>>
>
> It would be interesting to know what the issues with Open MPI are --
> hopefully they would be easy to resolve. Since there were two
> requests within a week for MPI_SUM/MPI_PROD and MPI_REAL*, I added
> support for the functionality in our release branch. It is also
> available in the 7.1.3a1 release I posted this afternoon.
>
> http://www.lam-mpi.org/beta/
>
>
> Brian
>
>
Brian: Thanks - I tried compiling 7.1.3a1 on osx and get this error:

gcc-3.3 -DHAVE_CONFIG_H -I. -I. -I../../share/include
-DLAM_TOTALVIEW_DLL=\"/sw/lammpi-xlf/lib/lam/liblam_totalview.so\"
-I../../share/include -DLAM_BUILDING=1 -D_REENTRANT -O3 -MT lamreduce.lo
-MD -MP -MF .deps/lamreduce.Tpo -c lamreduce.c -o lamreduce.o
lamreduce.c:34: warning: use of `long double' type; its size may change
in a future release
lamreduce.c:34: warning: (Long double usage is reported only once for
each file.
lamreduce.c:34: warning: To disable this warning, use -Wno-long-double.)
lamreduce.c: In function `lam_max':
lamreduce.c:315: error: `pfr16in' undeclared (first use in this function)
lamreduce.c:315: error: (Each undeclared identifier is reported only once
lamreduce.c:315: error: for each function it appears in.)
lamreduce.c:316: error: `pfr16out' undeclared (first use in this function)
lamreduce.c: In function `lam_min':
lamreduce.c:592: error: `pfr16in' undeclared (first use in this function)
lamreduce.c:593: error: `pfr16out' undeclared (first use in this function)
lamreduce.c: In function `lam_sum':
lamreduce.c:869: error: `pfr16in' undeclared (first use in this function)
lamreduce.c:870: error: `pfr16out' undeclared (first use in this function)
lamreduce.c: In function `lam_prod':
lamreduce.c:1155: error: `pfr16in' undeclared (first use in this function)
lamreduce.c:1156: error: `pfr16out' undeclared (first use in this function)

As for openmpi, I get a lot of messages like this

global_ssi(1441) malloc: *** Deallocation of a pointer not malloced:
0xbffe9e70; This could be a double free(), or free() called with the
middle of an allocated block; Try setting environment variable
MallocHelp to see tools to help debug
global_ssi(1441) malloc: *** Deallocation of a pointer not malloced:
0x1; This could be a double free(), or free() called with the middle of
an allocated block; Try setting environment variable MallocHelp to see
tools to help debug

and then it crashes randomly with a seg fault (no error message).

-Jeff

-- 
Jeffrey S. Whitaker         Phone : (303)497-6313
NOAA/OAR/CDC  R/PSD1        FAX   : (303)497-6449
325 Broadway                Boulder, CO, USA 80305-3328