<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=us-ascii">
<META content="MSHTML 6.00.6000.16608" name=GENERATOR></HEAD>
<BODY>
<DIV dir=ltr align=left><SPAN class=733455916-28022008><FONT face=Arial
color=#0000ff size=2>I'd say this mailing list is better for it right now.
Posting on the broader mpich-discuss might just distract us.</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=733455916-28022008><FONT face=Arial
color=#0000ff size=2></FONT></SPAN> </DIV>
<DIV dir=ltr align=left><SPAN class=733455916-28022008><FONT face=Arial
color=#0000ff size=2>Rajeev</FONT></SPAN></DIV><BR>
<BLOCKQUOTE
style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px solid; MARGIN-RIGHT: 0px">
<DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
<HR tabIndex=-1>
<FONT face=Tahoma size=2><B>From:</B> dcmf-bounces@lists.anl-external.org
[mailto:dcmf-bounces@lists.anl-external.org] <B>On Behalf Of </B>Bob
Cernohous<BR><B>Sent:</B> Wednesday, February 27, 2008 4:19 PM<BR><B>To:</B>
dcmf@lists.anl-external.org<BR><B>Subject:</B> Re: [dcmf] 32-bit ROMIO: What
is the best solution?<BR></FONT><BR></DIV>
<DIV></DIV><BR><TT><FONT size=2>bobc@us.ibm.com wrote on 02/08/2008 02:41:40
PM:</FONT></TT> <BR><TT><FONT size=2><BR>>We all realize there are issues
with MPI_Aint's in 32 bit implementations. </FONT></TT><BR><TT><FONT
size=2>>Using signed 32 bit addresses along with 64 bit offsets can result
in </FONT></TT><BR><TT><FONT size=2>>some pretty broken code. I've
reproduced several problems on </FONT></TT><BR><TT><FONT
size=2>>BGL/BGP/linux. The problems are most obvious with romio
files > 2G or </FONT></TT><BR><TT><FONT size=2>>virtual addresses >
2G.</FONT></TT> <BR><TT><FONT size=2><BR>Earlier this month I posted some
comments to this mailing list about </FONT></TT><BR><TT><FONT size=2>32-bit
MPICH/ROMIO/MPI_Aint issues (see above and the archive).
</FONT></TT> <BR><BR><TT><FONT size=2>We've started to work on the
problem. We're (optionally) making </FONT></TT><BR><TT><FONT
size=2>MPI_Aint a 64 bit signed value even when integers/pointers are 32 bit
</FONT></TT><BR><TT><FONT size=2>values.</FONT></TT> <BR><TT><FONT
size=2><BR>We'd like to start sending fix patches to the community for
comment.</FONT></TT> <BR><TT><FONT size=2><BR>However, since we are (in
parallel) contributing our BGP code back</FONT></TT> <BR><TT><FONT size=2>to
MPICH and trying to converge on a common code base of 1.0.7,</FONT></TT>
<BR><TT><FONT size=2>and because this problem applies to any 32-bit
platform,</FONT></TT> <BR><TT><FONT size=2>we thought we should move the
discussion to mpich-discuss to get</FONT></TT> <BR><TT><FONT size=2>the larger
community's comments. </FONT></TT> <BR><TT><FONT size=2><BR>It's
unlikely, at this point, that our patches would apply to MPICH2,</FONT></TT>
<BR><TT><FONT size=2>but that is our goal after the 1.0.7/BGP patch settles.
For now, we'd</FONT></TT> <BR><TT><FONT size=2>be sending these as RFC
(request for comment) patches.</FONT></TT> <BR><TT><FONT size=2><BR>Please let
us know if you agree with discussing this on mpich-discuss,</FONT></TT>
<BR><TT><FONT size=2>or if there is a better place.</FONT></TT> <BR><TT><FONT
size=2><BR>Bob Cernohous: (T/L 553)
507-253-6093<BR><BR>BobC@us.ibm.com<BR>IBM Rochester, Building 030-2(C335),
Department 61L<BR>3605 Hwy 52 North, Rochester, MN
55901-7829<BR><BR>> Chaos reigns within.<BR>> Reflect, repent, and
reboot.<BR>> Order shall return.<BR></BLOCKQUOTE></FONT></TT></BODY></HTML>