Error when doing Kpoint calculation in parallel

399 views
Skip to first unread message

Jacob Schulze

unread,
Sep 12, 2021, 11:07:25 AM9/12/21
to cp2k
Hey all,

I am having an issue where my calculation stops due to an error that occurs once the SCF loop starts. However, this only occurs if I ran the calculation over a Kpoint mesh on multiple processors. I can run the kpoint calculation serially just fine, and I can run the calculation in parallel without a kpoint mesh specified. 

I attached the input files and output file. Also the version of cp2k I am running is 8.0.

Anything I can do to stop this error from occuring? 

Here is the error message in full:
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.
{    0,    0}:  On entry to Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to DLASCL.

DSTEQR parameter number  -31 had an illegal value {    2,    2}:  On entry to 

DSTEQR parameter number  -31 had an illegal value 
{    1,    1}:  On entry to 
DSTEQR parameter number  -31 had an illegal value
LSCO.inp
LSCO2.out
LCO_SCAN5.xyz

hut...@chem.uzh.ch

unread,
Sep 15, 2021, 6:31:14 AM9/15/21
to cp...@googlegroups.com
Hi

I cannot reproduce your error, see my output. I don't think
we had any bug fixes recently in this part of the code, so my
best guess is that there is an inconsistency in your library
setup.

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Institut für Chemie E-mail: hut...@chem.uzh.ch
Universität Zürich
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----"'Jacob Schulze' via cp2k" <cp...@googlegroups.com> wrote: -----
To: "cp2k" <cp...@googlegroups.com>
From: "'Jacob Schulze' via cp2k" <cp...@googlegroups.com>
Date: 09/12/2021 05:07PM
Subject: [CP2K:15861] Error when doing Kpoint calculation in parallel
--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/cd011e96-edea-46db-b33d-4f18ab99b532n%40googlegroups.com.


[attachment "LSCO.inp" removed by Jürg Hutter/at/UZH]
[attachment "LSCO2.out" removed by Jürg Hutter/at/UZH]
[attachment "LCO_SCAN5.xyz" removed by Jürg Hutter/at/UZH]
LSCOx.out

Jacob Schulze

unread,
Sep 15, 2021, 10:27:11 AM9/15/21
to cp2k
Hey Juerg,

That did seem to be my problem. I switched the library I was using on the supercomputer, and now it works just fine. Thanks for the help!

Jacob
Reply all
Reply to author
Forward
0 new messages