trying to install; after build, get segmentation error on th  [SOLVED]

option, parallelism,...

Moderators: fgoudreault, mcote

Forum rules
Please have a look at ~abinit/doc/config/build-config.ac in the source package for detailed and up-to-date information about the configuration of Abinit 8 builds.
For a video explanation on how to build Abinit 7.x for Linux, please go to: http://www.youtube.com/watch?v=DppLQ-KQA68.
IMPORTANT: when an answer solves your problem, please check the little green V-like button on its upper-right corner to accept it.
Locked
MurrayDaw
Posts: 2
Joined: Tue May 31, 2016 8:01 pm

trying to install; after build, get segmentation error on th

Post by MurrayDaw » Tue May 31, 2016 8:11 pm

I did a "quickie" install on my MacBook Pro, using
./configure and
make

Seemed to compile and load OK, but when I ran the first test suggested in the on-line install notes
make test_fast

I got a segmentation fault.

So I went down into the test directory and did it manually:
./abinit < testin_fast.files > testin_fast_stdout.txt

which showed some steps before the segfault. I tried to attach the stdout here but it won't let me.
So I clipped the last part of the stdout here:

==== Info on pseudopotentials ====
Norm-conserving pseudopotentials
Number of pseudopotentials .. 1
Number of types of atoms .. 1
Scalar calculation (no spin-orbit term)
Nonlocal part applied using Legendre polynomials
Highest angular momentum +1 ....... 1
Max number of (l,n) components .. 1
Max number of (l,m,n) components .. 1

Info on the Q-grid used for form factors in spline form:
Number of q-points for radial functions ffspl .. 3001
Number of q-points for vlspl ................... 3001
vloc is computed in Reciprocal Space
model core charge treated in real-space

XC functional for type 1 is 1
Pseudo valence available: no
Pseudo number 1 read from ../../Psps_for_tests/01h.pspgth
Format or code 2
Generation date 960508
Content of first line: Goedecker-Teter-Hutter Wed May 8 14:27:44 EDT 1996

wfconv: 1 bands initialized randomly with npw= 544, for ikpt= 1
_setup2: Arith. and geom. avg. npw (full set) are 544.000 544.000
initro: for itypat= 1, take decay length= 0.6000,
initro: indeed, coreel= 0.0000, nval= 1 and densty= 0.0000E+00.

================================================================================
status : statnu, output_rate, shift_rate= 19 99 1
level,routine= 102 init
active(level)= 1
counter,actual_counter(level,1:2)= 0 0 0
List of active levels :
Active level number= 1
Active level number= 2
Active level number= 100

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x113ebf729
#1 0x113ebeaf5
#2 0x7fff90340529
#3 0x7fff9db6b93e
#4 0x10fcd4192
#5 0x10fcd43ec
#6 0x10f46d9ff
#7 0x10f0dd20a
#8 0x10f0e3e11
#9 0x10f0b604d
#10 0x10f0a4dfe
#11 0x10f02cf23
#12 0x10f007394
#13 0x10f00c130
#14 0x10efe6d29
#15 0x10efd11e0
#16 0x10fd3feee

User avatar
gmatteo
Posts: 291
Joined: Sun Aug 16, 2009 5:40 pm

Re: trying to install; after build, get segmentation error o  [SOLVED]

Post by gmatteo » Tue May 31, 2016 9:29 pm

Possible explanation:
Your Abinit is linked against the Blas implementation provided by Mac OSx (a.k.a. veclib)
The ABI of the veclib routines zdotc and cdotc is not compatible with Fortran and this causes nasty
segmentation faults at the beginning of the run.

Were this true, I would recommend passing

--enable-zdot-bugfix="yes"

to the configure script, reconfigure and finally make clean && make

If you have gdb installed on your Mac and Abinit has been compiled with "-g"
you can try to run testin_fast.file under the control of the debugger.
Use

gdb ./abinit

to load the executable and start gdb
Then, inside the gdb terminal, execute:

run < testin_fast.files

wait for the segmentation fault and post the output
M

MurrayDaw
Posts: 2
Joined: Tue May 31, 2016 8:01 pm

Re: trying to install; after build, get segmentation error o

Post by MurrayDaw » Tue May 31, 2016 10:49 pm

Thank you, gmatteo. That solved the problem exactly. Thank you also for being so quick to reply!

Locked