<meta http-equiv="Content-Type" content="text/html; charset=utf-8">I think this is the same issue reported on the devel list yesterday. If it is, Daniel Ibanez provided a patch that I've merged into my (completely unofficial fork on github) and will push to the official MPICH review repo tonight.<div><br></div><div>If you want to try and/or verify Daniel's<span></span> patch, see github/jeffhammond/mpich. The branch you want should be obvious from the name (I can't remember what it is off hand).</div><div><br></div><div>Best,<br><div><br></div><div>Jeff <br><br>On Sunday, January 10, 2016, VandeVondele Joost <<a href="mailto:joost.vandevondele@mat.ethz.ch">joost.vandevondele@mat.ethz.ch</a>> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>
<div style="direction:ltr;font-family:Tahoma;color:#000000;font-size:10pt">Hi,<br>
<br>
we upgraded from mpich 3.1.2 to 3.2 and our nightly builds under valgrind show a new valgrind warning, which I suppose originates from the mpi library, rather than our code (given the stack trace below and the fact that they are absent with 3.1):<br>
<br>
There are two slightly distinct traces (alltoall and alltoallv):<br>
<br>
==30165== Conditional jump or move depends on uninitialised value(s)<br>
==30165== at 0x6E00056: MPIR_Process_status (mpiimpl.h:4394)<br>
==30165== by 0x6E00056: MPIC_Waitall (helper_fns.c:774)<br>
==30165== by 0x6D5A853: MPIR_Alltoall_intra (alltoall.c:369)<br>
==30165== by 0x6D5B63C: MPIR_Alltoall (alltoall.c:564)<br>
==30165== by 0x6D5B6EF: MPIR_Alltoall_impl (alltoall.c:599)<br>
==30165== by 0x6D5BF07: PMPI_Alltoall (alltoall.c:722)<br>
==30165== by 0x6AFF0DC: PMPI_ALLTOALL (alltoallf.c:276)<br>
==30165== by 0x1D04699: __message_passing_MOD_mp_alltoall_i22 <br>
<br>
==30201== Conditional jump or move depends on uninitialised value(s)<br>
==30201== at 0x6E00056: MPIR_Process_status (mpiimpl.h:4394)<br>
==30201== by 0x6E00056: MPIC_Waitall (helper_fns.c:774)<br>
==30201== by 0x6D5C4EA: MPIR_Alltoallv_intra (alltoallv.c:194)<br>
==30201== by 0x6D5CEC9: MPIR_Alltoallv (alltoallv.c:339)<br>
==30201== by 0x6D5CFAA: MPIR_Alltoallv_impl (alltoallv.c:376)<br>
==30201== by 0x6D5D8CA: PMPI_Alltoallv (alltoallv.c:527)<br>
==30201== by 0x6AFF13E: PMPI_ALLTOALLV (alltoallvf.c:276)<br>
==30201== by 0x1CCE180: __message_passing_MOD_mp_alltoall_z22v (message_passing_z.f90:223)<br>
<br>
Is this issue already known/fixed ?<br>
<br>
I saw there was a somewhat related change (which is part of 3.2):<br>
<br>
<a href="https://trac.mpich.org/projects/mpich/changeset/be7c17b3177ccb0ede49761b44a438952360b061" target="_blank">https://trac.mpich.org/projects/mpich/changeset/be7c17b3177ccb0ede49761b44a438952360b061</a><br>
<br>
but maybe that change was not fully complete ?<br>
<br>
Best regards,<br>
<br>
Joost VandeVondele<br>
<div><br>
<div style="font-family:Tahoma;font-size:13px">
<div style="font-family:Tahoma;font-size:13px">
<div style="font-family:Tahoma;font-size:13px">
<hr style="width:100%;min-height:2px">
Prof. Dr. Joost VandeVondele<br>
Nanoscale Simulations<br>
<br>
ETH Zurich <br>
HIT G 41.4<span style="font-family:monospace"><br>
</span>Wolfgang-Pauli-Strasse 27<br>
8093 Zurich, Switzerland <br>
<br>
<a href="javascript:_e(%7B%7D,'cvml','Joost.VandeVondele@mat.ethz.ch');" target="_blank">Joost.VandeVondele@mat.ethz.ch</a><br>
<a href="http://www.nanosim.mat.ethz.ch/" target="_blank">http://www.nanosim.mat.ethz.ch/</a><br>
Tel.: +41 (0)44 6338451<br>
<hr style="width:100%;min-height:2px">
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote></div></div><br><br>-- <br>Jeff Hammond<br><a href="mailto:jeff.science@gmail.com" target="_blank">jeff.science@gmail.com</a><br><a href="http://jeffhammond.github.io/" target="_blank">http://jeffhammond.github.io/</a><br>