<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">
Actually, afaik, Win_fence is not the right synchronization method for you approach. As Halim said, you should use
<div class="">passive-target synchronization here. Fence guarantees you that all RMA operations on win originating at a given</div>
<div class="">process and started before the fence call will complete at that process before the fence call return - so it only</div>
<div class="">synchronizes RMA calls - but not multiple accesses to the same are. </div>
<div class=""><br class="">
</div>
<div class="">However, in you example that does NOT avoid that two processes access<i class=""> a</i> at the same time</div>
<div class="">- that why you get your race-condition.</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
<div class="">Lena </div>
<div class=""><br class="">
<div>
<blockquote type="cite" class="">
<div class="">On Apr 13, 2016, at 3:08 PM, Halim Amer <<a href="mailto:aamer@anl.gov" class="">aamer@anl.gov</a>> wrote:</div>
<br class="Apple-interchange-newline">
<div class="">
<div class="">Win_sync can only be used within passive target epochs. You could replace the win_fences with win_locks.<br class="">
<br class="">
--Halim<br class="">
<br class="">
On 4/13/16 11:28 AM, C.Friedrich wrote:<br class="">
<blockquote type="cite" class="">Thanks for the suggestion! Adding MPI_WIN_SYNC in this way:<br class="">
do i = 1,10000<br class="">
call mpe_win_fence(a,0,Nwin,err)<br class="">
a = a + 1<br class="">
call mpi_win_sync(Nwin,err)<br class="">
call mpe_win_fence(a,0,Nwin,err)<br class="">
enddo<br class="">
does not work. The compiled executable either issues the error<br class="">
Fatal error in MPI_Win_sync: Wrong synchronization of RMA calls , error<br class="">
stack:<br class="">
MPI_Win_sync(115)..: MPI_Win_sync(win=0xa0000000) failed<br class="">
MPID_Win_sync(1805): Wrong synchronization of RMA calls<br class="">
or simply dies (without writing the numbers). Perhaps I use it<br class="">
incorrectly...<br class="">
<br class="">
Christoph<br class="">
<br class="">
<br class="">
<br class="">
------------------------------------------------------------------------------------------------<br class="">
<br class="">
------------------------------------------------------------------------------------------------<br class="">
<br class="">
Forschungszentrum Juelich GmbH<br class="">
52425 Juelich<br class="">
Sitz der Gesellschaft: Juelich<br class="">
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498<br class="">
Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher<br class="">
Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender),<br class="">
Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,<br class="">
Prof. Dr. Sebastian M. Schmidt<br class="">
------------------------------------------------------------------------------------------------<br class="">
<br class="">
------------------------------------------------------------------------------------------------<br class="">
<br class="">
<br class="">
_______________________________________________<br class="">
discuss mailing list <a href="mailto:discuss@mpich.org" class="">discuss@mpich.org</a><br class="">
To manage subscription options or unsubscribe:<br class="">
<a href="https://lists.mpich.org/mailman/listinfo/discuss" class="">https://lists.mpich.org/mailman/listinfo/discuss</a><br class="">
</blockquote>
_______________________________________________<br class="">
discuss mailing list <a href="mailto:discuss@mpich.org" class="">discuss@mpich.org</a><br class="">
To manage subscription options or unsubscribe:<br class="">
<a href="https://lists.mpich.org/mailman/listinfo/discuss" class="">https://lists.mpich.org/mailman/listinfo/discuss</a><br class="">
</div>
</div>
</blockquote>
</div>
<br class="">
</div>
</body>
</html>