VASP Segmentation Fault Seemingly due to Vacuum Width?

Problems running VASP: crashes, internal errors, "wrong" results.

Moderators: Global Moderator, Moderator

Post Reply
Message
Author
meb7pz
Newbie
Newbie
Posts: 2
Joined: Sun May 12, 2019 9:29 pm
License Nr.: 5-2625

VASP Segmentation Fault Seemingly due to Vacuum Width?

#1 Post by meb7pz » Tue Mar 03, 2020 12:29 am

I have multiple systems with the following Bravais lattice and k-mesh:
From POSCAR:
11.26488000 0.00000000 0.00000000
-5.63244000 9.75567225 0.00000000
0.00000000 0.00000000 30.00000000

a.k.a.

A: 11.26488
B: 11.26488
C: 30
alpha:90
beta: 90
gamma: 60 or 120

From KPOINTS:
Automatic mesh
0
Gamma
8 8 1
0. 0. 0.

I am modeling a bulk system for Pt(111) (64-80 metal atoms in total), as well as small adsorbates (<10 atoms). This run fails with the above lattice and ENCUT=400 (C ENMAX), but tends to run when the z-dimension is reduced to 20 or 10 A or when ENCUT is set to 230.283 (Pt ENMAX). When I try minimizing the the same unit cell with only the adsorbate present it takes reducing the z-dimension of the unit cell to <10 A before the calculation might run. I do change the POTCAR to reflect the change in elements. (INCAR, example POSCAR, KPOINTS, and error_file below)

Some hunting online has shown that reducing the vacuum size does help reduce the likelihood of SEGSIEV errors, but I do not see any clear explanations as to why this is the case, nor can I find the relationship between ENCUT and unit cell size.

As a result, I am looking for an explanation for why reducing the size of my unit cell is allowing my calculations to run. Is there a scaling law I should be following to ensure I do not get this error in the future? I have looked on this forum and in the VASP manual for more information, but have yet to find an explanation. If this phenomena is discussed elsewhere, I would appreciate someone pointing me in that direction.

Thanks,
Mark
##########################################################
Working INCAR
##########################################################
PREC = Accurate
ENCUT = 230.283
IBRION = 3
NSW = 250
EDIFF = 1.0e-04
EDIFFG = -0.01
ALGO = Normal (blocked Davidson)
IVDW = 11
VDW_S6 = 1.0
VDW_S8 = 0.722
VDW_SR = 1.217
#SYMPREC = 10e-4

# Defaults
NWRITE = 2
ISIF = 2
NELMIN = 2
NBLOCK = 1
NELM = 60
ISPIN = 1
INIWAV = 1
ISTART = 0
ICHARG = 2
LWAVE = .FALSE.
LCHARG = .FALSE.
ADDGRID = .FALSE.
LHYPERFINE = .FALSE.
ISMEAR = 1
SIGMA = 0.2
LREAL = .FALSE.
##########################################################
Working POSCAR
##########################################################
(Pt)64 (P1) ~ Pt 1 1 1 surface_4x4x1
1.00000000000000
11.2648799999999998 0.0000000000000000 0.0000000000000000
-5.6324399999999999 9.7556722499999999 0.0000000000000000
0.0000000000000000 0.0000000000000000 30.0000000000000000
Pt
64
Selective dynamics
Direct
0.0021286799999984 0.0013231599999983 0.2249378682276826 T T T
0.0854620100000005 0.1679898200000025 0.2999680875985232 T T T
0.0854620100000005 0.1679898200000025 0.0749533500000013 F F F
0.1687953500000035 0.0846564900000004 0.1516010100000003 F F F
0.0021286799999984 0.2513231599999983 0.2249378682276826 T T T
0.0854620100000005 0.4179898200000025 0.2999680875985232 T T T
0.0854620100000005 0.4179898200000025 0.0749533500000013 F F F
0.1687953500000035 0.3346564900000004 0.1516010100000003 F F F
0.0021286799999984 0.5013231599999983 0.2249378682276826 T T T
0.0854620100000005 0.6679898200000025 0.2999680875985232 T T T
0.0854620100000005 0.6679898200000025 0.0749533500000013 F F F
0.1687953500000035 0.5846564900000004 0.1516010100000003 F F F
0.0021286799999984 0.7513231599999983 0.2249378682276826 T T T
0.0854620100000005 0.9179898200000025 0.2999680875985232 T T T
0.0854620100000005 0.9179898200000025 0.0749533500000013 F F F
0.1687953500000035 0.8346564900000004 0.1516010100000003 F F F
0.2521286799999984 0.0013231599999983 0.2249378682276826 T T T
0.3354620100000005 0.1679898200000025 0.2999680875985232 T T T
0.3354620100000005 0.1679898200000025 0.0749533500000013 F F F
0.4187953500000035 0.0846564900000004 0.1516010100000003 F F F
0.2521286799999984 0.2513231599999983 0.2249378682276826 T T T
0.3354620100000005 0.4179898200000025 0.2999680875985232 T T T
0.3354620100000005 0.4179898200000025 0.0749533500000013 F F F
0.4187953500000035 0.3346564900000004 0.1516010100000003 F F F
0.2521286799999984 0.5013231599999983 0.2249378682276826 T T T
0.3354620100000005 0.6679898200000025 0.2999680875985232 T T T
0.3354620100000005 0.6679898200000025 0.0749533500000013 F F F
0.4187953500000035 0.5846564900000004 0.1516010100000003 F F F
0.2521286799999984 0.7513231599999983 0.2249378682276826 T T T
0.3354620100000005 0.9179898200000025 0.2999680875985232 T T T
0.3354620100000005 0.9179898200000025 0.0749533500000013 F F F
0.4187953500000035 0.8346564900000004 0.1516010100000003 F F F
0.5021286799999984 0.0013231599999983 0.2249378682276826 T T T
0.5854620100000005 0.1679898200000025 0.2999680875985232 T T T
0.5854620100000005 0.1679898200000025 0.0749533500000013 F F F
0.6687953500000035 0.0846564900000004 0.1516010100000003 F F F
0.5021286799999984 0.2513231599999983 0.2249378682276826 T T T
0.5854620100000005 0.4179898200000025 0.2999680875985232 T T T
0.5854620100000005 0.4179898200000025 0.0749533500000013 F F F
0.6687953500000035 0.3346564900000004 0.1516010100000003 F F F
0.5021286799999984 0.5013231599999983 0.2249378682276826 T T T
0.5854620100000005 0.6679898200000025 0.2999680875985232 T T T
0.5854620100000005 0.6679898200000025 0.0749533500000013 F F F
0.6687953500000035 0.5846564900000004 0.1516010100000003 F F F
0.5021286799999984 0.7513231599999983 0.2249378682276826 T T T
0.5854620100000005 0.9179898200000025 0.2999680875985232 T T T
0.5854620100000005 0.9179898200000025 0.0749533500000013 F F F
0.6687953500000035 0.8346564900000004 0.1516010100000003 F F F
0.7521286799999984 0.0013231599999983 0.2249378682276826 T T T
0.8354620100000005 0.1679898200000025 0.2999680875985232 T T T
0.8354620100000005 0.1679898200000025 0.0749533500000013 F F F
0.9187953500000035 0.0846564900000004 0.1516010100000003 F F F
0.7521286799999984 0.2513231599999983 0.2249378682276826 T T T
0.8354620100000005 0.4179898200000025 0.2999680875985232 T T T
0.8354620100000005 0.4179898200000025 0.0749533500000013 F F F
0.9187953500000035 0.3346564900000004 0.1516010100000003 F F F
0.7521286799999984 0.5013231599999983 0.2249378682276826 T T T
0.8354620100000005 0.6679898200000025 0.2999680875985232 T T T
0.8354620100000005 0.6679898200000025 0.0749533500000013 F F F
0.9187953500000035 0.5846564900000004 0.1516010100000003 F F F
0.7521286799999984 0.7513231599999983 0.2249378682276826 T T T
0.8354620100000005 0.9179898200000025 0.2999680875985232 T T T
0.8354620100000005 0.9179898200000025 0.0749533500000013 F F F
0.9187953500000035 0.8346564900000004 0.1516010100000003 F F F
##########################################################
Non-functioning INCAR
##########################################################
PREC = Accurate
ENCUT = 400
IBRION = 3
NSW = 250
EDIFF = 1.0e-04
EDIFFG = -0.01
ALGO = Normal (blocked Davidson)
IVDW = 11
VDW_S6 = 1.0
VDW_S8 = 0.722
VDW_SR = 1.217
#SYMPREC = 10e-4

# Defaults
NWRITE = 2
ISIF = 2
NELMIN = 2
NBLOCK = 1
NELM = 60
ISPIN = 1
INIWAV = 1
ISTART = 0
ICHARG = 2
LWAVE = .FALSE.
LCHARG = .FALSE.
ADDGRID = .FALSE.
LHYPERFINE = .FALSE.
ISMEAR = 1
SIGMA = 0.2
LREAL = .FALSE.
##########################################################
Non-functioning POSCAR
##########################################################
(C)1 (H)4 (P1)
1.00000000000000
11.2648799999999998 0.0000000000000000 0.0000000000000000
-5.6324399999999999 9.7556722499999999 0.0000000000000000
0.0000000000000000 0.0000000000000000 30.0000000000000000
C H
1 4
Selective dynamics
Direct
0.5194936700000000 0.3147349800000000 0.5262534900000000 T T T
0.6049203200000000 0.2087695000000000 0.5262534600000000 T T T
0.5254483200000000 0.3677169800000000 0.5560955700000000 T T T
0.5254481400000000 0.3677170800000000 0.4964114100000000 T T T
0.4221579100000000 0.3147363400000000 0.5262535200000000 T T T
##########################################################
error_file
##########################################################
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
vasp_std 00000000016D0FC5 Unknown Unknown Unknown
vasp_std 00000000016CED87 Unknown Unknown Unknown
vasp_std 000000000164EFD4 Unknown Unknown Unknown
vasp_std 000000000164EDE6 Unknown Unknown Unknown
vasp_std 00000000015F0C46 Unknown Unknown Unknown
vasp_std 00000000015F69C0 Unknown Unknown Unknown
libpthread.so.0 00007FF745CD25F0 Unknown Unknown Unknown
vasp_std 00000000013F6DEF Unknown Unknown Unknown
vasp_std 00000000013F6181 Unknown Unknown Unknown
vasp_std 000000000086B0B8 Unknown Unknown Unknown
vasp_std 0000000000E77BA1 Unknown Unknown Unknown
vasp_std 000000000147B7BE Unknown Unknown Unknown
vasp_std 0000000001458BE6 Unknown Unknown Unknown
vasp_std 000000000040A06E Unknown Unknown Unknown
libc.so.6 00007FF745615505 Unknown Unknown Unknown
vasp_std 0000000000409F79 Unknown Unknown Unknown
srun: error: udc-ba35-16c7: task 0: Exited with exit code 174
srun: Terminating job step 9206835.0
slurmstepd: error: *** STEP 9206835.0 ON udc-ba35-16c7 CANCELLED AT 2020-03-02T19:08:34 ***
forrtl: error (78): process killed (SIGTERM)
Image PC Routine Line Source
vasp_std 00000000016D0FC5 Unknown Unknown Unknown
vasp_std 00000000016CED87 Unknown Unknown Unknown
vasp_std 000000000164EFD4 Unknown Unknown Unknown
vasp_std 000000000164EDE6 Unknown Unknown Unknown
vasp_std 00000000015F0C46 Unknown Unknown Unknown
vasp_std 00000000015F698E Unknown Unknown Unknown
libpthread.so.0 00007F287AEF35F0 Unknown Unknown Unknown
libmpi.so.12 00007F287B642D61 Unknown Unknown Unknown
libmpi.so.12 00007F287B7AA1E0 Unknown Unknown Unknown
libmpi.so.12 00007F287B6277AE Unknown Unknown Unknown
libmpi.so.12 00007F287B62DB67 Unknown Unknown Unknown
libmpi.so.12 00007F287B62D1B3 Unknown Unknown Unknown
libmpifort.so.12 00007F287C1F3924 Unknown Unknown Unknown
vasp_std 0000000000413EF4 Unknown Unknown Unknown
vasp_std 000000000086B101 Unknown Unknown Unknown
vasp_std 0000000000E77BA1 Unknown Unknown Unknown
vasp_std 000000000147B7BE Unknown Unknown Unknown
vasp_std 0000000001458BE6 Unknown Unknown Unknown
vasp_std 000000000040A06E Unknown Unknown Unknown
libc.so.6 00007F287A836505 Unknown Unknown Unknown
vasp_std 0000000000409F79 Unknown Unknown Unknown
...

alex
Hero Member
Hero Member
Posts: 567
Joined: Tue Nov 16, 2004 2:21 pm
License Nr.: 5-67
Location: Germany

Re: VASP Segmentation Fault Seemingly due to Vacuum Width?

#2 Post by alex » Mon Mar 09, 2020 8:05 am

Hi meb7pz,

my number one guess on basis of your error 'forrtl: error (78): process killed (SIGTERM)' is you might have a memory issue, because
- if your system get's smaller, it works (less vacuum)
- lower cut-off energy, it works
- your program becomes killed by the operating system: a very high penalty: ups.

Hth,

alex

meb7pz
Newbie
Newbie
Posts: 2
Joined: Sun May 12, 2019 9:29 pm
License Nr.: 5-2625

Re: VASP Segmentation Fault Seemingly due to Vacuum Width?

#3 Post by meb7pz » Mon Mar 16, 2020 8:33 pm

Thanks Alex,

I am currently running with 6000 megabytes per cpu (I am running off a SLURM workload manager). I upped the memory per cpu to 9000 megabytes, but the error persists. If I decrease the x and y coordinates of the unit cell to:
5.63244000 0.00000000 0.00000000
-2.81622000 4.87783613 0.00000000
0.00000000 0.00000000 30.69487000

Then I can get the calculation to run, but for larger adsorbates (e.g. C5H12) such a small unit cell won't cut it (I am looking for initial sticking probabilities and want to omit coverage effects as much as possible). Additionally, I dropped the KPOINTS for the original unit cell down to 2 2 1 and the calculation still failed with 9000 megabytes. Plus, if I increase the ENCUT by even 10 eV (i.e. 240 eV) the calculation also fails.

Is there any way for me to calculate how much memory I should be using to get these larger calculations to run?

Thank you,

Mark

henrique_miranda
Global Moderator
Global Moderator
Posts: 414
Joined: Mon Nov 04, 2019 12:41 pm
Contact:

Re: VASP Segmentation Fault Seemingly due to Vacuum Width?

#4 Post by henrique_miranda » Thu Mar 19, 2020 11:07 am

Is there any way for me to calculate how much memory I should be using to get these larger calculations to run?
Currently this is not possible.

I recommend that you look into how much memory your current working job (the one with less vaccum) is using.
Then you can slowly increase the vaccum and see how much the memory is increasing.
This will give you an idea of how much memory your job will need.

PS: 8x8x1 KPOINS seems exaggerated for a surface supercell calculation, please check how the quantity of interest changes when reducing the k-points.

jianan_zhang
Newbie
Newbie
Posts: 2
Joined: Fri Mar 06, 2020 4:27 pm

Re: VASP Segmentation Fault Seemingly due to Vacuum Width?

#5 Post by jianan_zhang » Thu Mar 19, 2020 11:39 pm

Hi,

I found the same error as you do when I try to run vasp on a cluster, it is possible it's due to stack size limit. If it's due to that reason, you can run "ulimit -s unlimited" in the terminal then problem will be solved, or you can add this line in ~/.bashrc so this line ran every time you have a new terminal.

Post Reply