[QE-users] MPI in qe-6.5

Fri Dec 20 17:47:48 CET 2019

Hi Ian,

I was able to get this crash when I compiled w/o scalapack (nor elpa, etc.). Assuming you’re also running w/o scalapack, try editing LAXlib/ptoolkit.f90. Starting at line 4212 you should have:

    ! split communicator is present and must be freed on all processors
    CALL mpi_comm_free( col_comm, ierr )
    IF( ierr /= 0 ) &

Change it to:

    ! split communicator is present and must be freed on all processors
    IF( col_comm /= MPI_COMM_NULL ) THEN
      CALL mpi_comm_free( col_comm, ierr )
      IF( ierr /= 0 ) &
            CALL lax_error__( " pdtrtri ", " in mpi_comm_free 25 ", ABS( ierr ) )
    END IF

And see if that fixes it for you.

John

From: users <users-bounces at lists.quantum-espresso.org> on behalf of Ian Shuttleworth <shuttleworth.ian at gmail.com>
Reply-To: Quantum ESPRESSO users Forum <users at lists.quantum-espresso.org>
Date: Friday, December 20, 2019 at 10:59 AM
To: Quantum ESPRESSO users Forum <users at lists.quantum-espresso.org>
Subject: Re: [QE-users] MPI in qe-6.5

I've attached "test.in<https://gcc01.safelinks.protection.outlook.com/?url=http%3A%2F%2Ftest.in&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805045449&sdata=6MhdnfIg9tQ2YmKTC6bye%2BaAaGnYStkcSCZ3flJgRrc%3D&reserved=0>" the input file - the pseudos just come directly from the QE web-site

With thanks

Ian

On Fri, Dec 20, 2019 at 3:47 PM Pietro Delugas <pdelugas at sissa.it<mailto:pdelugas at sissa.it>> wrote:

Dear Ian

I just compiled pw with gcc-4.8.5 and openmpi 1.10.7 and the test-suite tests are passing, so either it is a problem in someway related to your input, to some issue related to the way you compiled the program or to some feature of you system other than the compiler and the mpi library.

kind regards - Pietro

Could you send you input
On 20/12/19 16:23, Ian Shuttleworth wrote:
Dear all

I encounter MPI communication errors when running 6.5 compiled both with gcc-4.8.5 and openMPI 1.10.7, and also - on another HPC - compiled with gcc 6.4.0 and open  OpenMPI 2.1.2.

The code starts correctly with statements:

     This program is part of the open-source Quantum ESPRESSO suite
     for quantum simulation of materials; please cite
         "P. Giannozzi et al., J. Phys.:Condens. Matter 21 395502 (2009);
         "P. Giannozzi et al., J. Phys.:Condens. Matter 29 465901 (2017);
          URL http://www.quantum-espresso.org<https://gcc01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.quantum-espresso.org&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805055443&sdata=0gEwitsOMxujynLItOAJe4Q8n3AUD97ZuXiEzjrMvd0%3D&reserved=0>",
     in publications or presentations arising from this work. More details at
     http://www.quantum-espresso.org/quote<https://gcc01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.quantum-espresso.org%2Fquote&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805055443&sdata=Uxu67hdAKo%2Fh2TGa%2FXA9EWLXx7v0QDFTfV07marSZX8%3D&reserved=0>
     Parallel version (MPI), running on    16 processors

The execution stops with the statement:

     Starting wfcs are  192 randomized atomic wfcs

and the following error messages then appear:

[node44:20988] *** An error occurred in MPI_Comm_free
[node44:20988] *** reported by process [140653000785921,140724603453442]
[node44:20988] *** on communicator MPI_COMM_WORLD
[node44:20988] *** MPI_ERR_COMM: invalid communicator
[node44:20988] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[node44:20988] ***    and potentially your MPI job)

Compiling 6.4.1 using exactly the same gcc/openmpi's doesn't produce the same problem, and the execution in fact completes without error. So my question is: what differences are there in the MPI between 6.5 and 6.4.1, and are there any 'tweaks' that could be applied to the compile script to remove the problems I'm seeing with 6.5?

With thanks

Ian Shuttleworth
(Nottingham Trent University)

_______________________________________________

Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso<https://gcc01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.max-centre.eu%2Fquantum-espresso&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805065437&sdata=1RmTscxQZflZOy0rA2mt1B%2FuDIIYt79fUX0ZN5AnOxs%3D&reserved=0>)

users mailing list users at lists.quantum-espresso.org<mailto:users at lists.quantum-espresso.org>

https://lists.quantum-espresso.org/mailman/listinfo/users<https://gcc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fmailman%2Flistinfo%2Fusers&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805065437&sdata=Byr38ph91oRMhfIUZkMGB0vosBDrKuqtZb%2Fc5CN%2F%2FXY%3D&reserved=0>
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso<https://gcc01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.max-centre.eu%2Fquantum-espresso&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805075434&sdata=%2BhVcfRsTvDVMzSRQwthUjUk7bTu1UGhA3WmxnnSJ%2Bh4%3D&reserved=0>)
users mailing list users at lists.quantum-espresso.org<mailto:users at lists.quantum-espresso.org>
https://lists.quantum-espresso.org/mailman/listinfo/users<https://gcc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fmailman%2Flistinfo%2Fusers&data=02%7C01%7Cjohn.vinson%40nist.gov%7Cc07d9fa036e5429e021908d785659d51%7C2ab5d82fd8fa4797a93e054655c61dec%7C1%7C1%7C637124543805075434&sdata=L5bIozKD3Aqs8Z1Y%2B7y2hBlZxLBMDIqCxANwcJZaONI%3D&reserved=0>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20191220/f6bcc1b3/attachment.html>