<div dir="ltr">I guess Jacob already used MPICH, since MPIDI_CH3_EagerContigShortSend() is from MPICH. <div><br clear="all"><div><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr">--Junchao Zhang</div></div></div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Jun 2, 2020 at 9:38 AM Satish Balay via petsc-dev <<a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">use --download-mpich for valgrind.<br>
<br>
<a href="https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind" rel="noreferrer" target="_blank">https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind</a><br>
<br>
Satish<br>
<br>
On Tue, 2 Jun 2020, Karl Rupp wrote:<br>
<br>
> Hi Jacob,<br>
> <br>
> the recommendation in the past was to use MPICH as it is (was?)<br>
> valgrind-clean. Which MPI do you use? OpenMPI used to have these kinds of<br>
> issues. (My information might be outdated)<br>
> <br>
> Best regards,<br>
> Karli<br>
> <br>
> On 6/2/20 2:43 AM, Jacob Faibussowitsch wrote:<br>
> > Hello All,<br>
> > <br>
> > TL;DR: valgrind always complains about "Syscall param write(buf) points to<br>
> > uninitialised byte(s)” for a LOT of MPI operations in petsc code, making<br>
> > debugging using valgrind fairly annoying since I have to sort through a ton<br>
> > of unrelated stuff. I have built valgrind from source, used apt install<br>
> > valgrind, apt install valgrind-mpi to no avail.<br>
> > <br>
> > I am using valgrind from docker. Dockerfile is attached below as well. I<br>
> > have been unsuccessfully trying to resolve these local valgrind errors, but<br>
> > I am running out of ideas. Googling the issue has also not provided entirely<br>
> > applicable solutions. Here is an example of the error:<br>
> > <br>
> > $ make -f gmakefile test VALGRIND=1<br>
> > ...<br>
> > #==54610== Syscall param write(buf) points to uninitialised byte(s)<br>
> > #==54610== at 0x6F63317: write (write.c:26)<br>
> > #==54610== by 0x9056AC9: MPIDI_CH3I_Sock_write (in <br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x9059FCD: MPIDI_CH3_iStartMsg (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x903F298: MPIDI_CH3_EagerContigShortSend (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x9049479: MPID_Send (in /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8FC9B2A: MPIC_Send (in /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8F86F2E: MPIR_Bcast_intra_binomial (in <br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8EE204E: MPIR_Bcast_intra_auto (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8EE21F4: MPIR_Bcast_impl (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8F887FB: MPIR_Bcast_intra_smp (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8EE206E: MPIR_Bcast_intra_auto (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8EE21F4: MPIR_Bcast_impl (in<br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x8EE2A6F: PMPI_Bcast (in /usr/local/lib/libmpi.so.12.1.8)<br>
> > #==54610== by 0x4B377B8: PetscOptionsInsertFile (options.c:525)<br>
> > #==54610== by 0x4B39291: PetscOptionsInsert (options.c:672)<br>
> > #==54610== by 0x4B5B1EF: PetscInitialize (pinit.c:996)<br>
> > #==54610== by 0x10A6BA: main (ex9.c:75)<br>
> > #==54610== Address 0x1ffeffa944 is on thread 1's stack<br>
> > #==54610== in frame #3, created by MPIDI_CH3_EagerContigShortSend (???:)<br>
> > #==54610== Uninitialised value was created by a stack allocation<br>
> > #==54610== at 0x903F200: MPIDI_CH3_EagerContigShortSend (in <br>
> > /usr/local/lib/libmpi.so.12.1.8)<br>
> > <br>
> > There are probably 20 such errors every single time, regardless of what code<br>
> > is being run. I have tried using apt install valgrind, apt install<br>
> > valgrind-mpi, and building valgrind from source:<br>
> > <br>
> > # VALGRIND<br>
> > WORKDIR /<br>
> > RUN git clone git://<a href="http://sourceware.org/git/valgrind.git" rel="noreferrer" target="_blank">sourceware.org/git/valgrind.git</a><br>
> > WORKDIR /valgrind<br>
> > RUN git pull<br>
> > RUN ./autogen.sh<br>
> > RUN ./configure --with-mpicc=/usr/local/bin/mpicc<br>
> > RUN make -j 5<br>
> > RUN make install<br>
> > <br>
> > None of the those approaches lead to these errors disappearing. Perhaps I am<br>
> > missing some funky MPI args?<br>
> > <br>
> > Best regards,<br>
> > <br>
> > Jacob Faibussowitsch<br>
> > (Jacob Fai - booss - oh - vitch)<br>
> > Cell: (312) 694-3391<br>
> > <br>
> > <br>
> <br>
> <br>
</blockquote></div>