Hello everyone, I think I found some bugs for the vasp 5.3.3.
I are running a tensile calculation for Diamond and a memory corruption happened when I put a small strain (about 5%) on it. I am very confused for the same calculation can be done with vasp 4.6. Besides, if I run it on one core, without mpi, it can finished as well. This bugs only happen when I use openmpi to run vasp 5.2 and vasp 5.3.I have tried to compile vasp with ifort 9 and ifort 11 but none of them work.
Here is my error message:
*** glibc detected *** vasp53_mpi_zz: malloc(): memory corruption: 0x00002aaaac6aa700 ***
======= Backtrace: =========
/lib64/libc.so.6[0x2ab0c7eed31e]
/lib64/libc.so.6[0x2ab0c7eef3b4]
/lib64/libc.so.6(malloc+0x86)[0x2ab0c7ef0766]
/public/software/openmpi-1.4.2/lib/openmpi/mca_coll_tuned.so[0x2ab0ca880572]
/public/software/openmpi-1.4.2/lib/openmpi/mca_coll_tuned.so[0x2ab0ca87e629]
/public/software/openmpi-1.4.2/lib/libmpi.so.0(PMPI_Allreduce+0x77)[0x2ab0c747997f]
/public/software/openmpi-1.4.2/lib/libmpi_f77.so.0(pmpi_allreduce+0x11b)[0x2ab0c731145b]
vasp53_mpi_zz(m_sum_d_+0xa7)[0x4d632d]
vasp53_mpi_zz(mrg_grid_rl_plane_+0x16e)[0x517c88]
vasp53_mpi_zz(fileio_mp_outchg_+0x5d5)[0x83a553]
vasp53_mpi_zz(MAIN__+0x486f8)[0x4c6df8]
vasp53_mpi_zz(main+0x32)[0x47e6ea]
/lib64/libc.so.6(__libc_start_main+0xf4)[0x2ab0c7e9f184]
vasp53_mpi_zz(calloc+0x259)[0x47e629]
======= Memory map: ========
Here is the last part of my running log
3 F= -.72352209E+02 E0= -.72352209E+02 d E =-.116974E-03
curvature: -1.23 expect dE=-0.287E-07 dE for cont linesearch -0.287E-07
trial: gam=-0.01549 g(F)= 0.699E-12 g(S)= 0.234E-07 ort = 0.149E-05 (trialstep = 0.129E+01)
search vector abs. value= 0.107E-09
This question is killing me. Please help and if you need further information just ask.
memory corruption
Moderators: Global Moderator, Moderator
-
- Newbie
- Posts: 2
- Joined: Fri Mar 22, 2013 7:27 am
memory corruption
Last edited by ppphysics on Thu May 09, 2013 5:43 pm, edited 1 time in total.
memory corruption
I encountered the same situation
Last edited by yunguo on Sun Mar 23, 2014 4:16 pm, edited 1 time in total.
-
- Hero Member
- Posts: 585
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
memory corruption
Hm, have you got ECC Memory installed (This is the one which recognizes bit errors)?
If not, try to run a memory checker.
If yes, search your system log for hints.
Cheers,
alex
<span class='smallblacktext'>[ Edited Mon Mar 24 2014, 07:54AM ]</span>
If not, try to run a memory checker.
If yes, search your system log for hints.
Cheers,
alex
<span class='smallblacktext'>[ Edited Mon Mar 24 2014, 07:54AM ]</span>
Last edited by alex on Mon Mar 24, 2014 7:54 am, edited 1 time in total.