Hello everyone,
Using Quantum espresso , I tried for Na3Bi calculations using Full relativistic Pseudopotential but everytime Iam getting error as "Unable to allocate memory" . I used every trick like reducing mixing_ndim, diagonalization=cg. but still Iam getting the same.
For Na: suggested ecut and rho are 66, 323 and for Bi is 45 and 456.
So to avoid this error I used lower ecut values as 30 and 300 for Na3Bi but still Iam getting the same error.
This is the input file
&control
calculation='scf',
restart_mode='from_scratch',
outdir='./tmp',
prefix='na3bi',
pseudo_dir='../pseudo/',
/
&system
ibrav=4,
celldm(1)=10.28, celldm(3)=1.773
nat=3,
ntyp=2,
ecutwfc = 30,
ecutrho=300,
lspinorb=.true.
noncolin=.true.
starting_magnetization(1) = 0.3
angle1(1) = 90.0
angle2(1) = 0.0
starting_magnetization(2) = 0.3
angle1(2) = 90.0
angle2(2) = 0.0
occupations = 'smearing',
smearing = 'marzari-vanderbilt',
degauss = 0.04
/
&electrons
mixing_beta=0.3,
diagonalization='cg'
mixing_ndim=4
/
ATOMIC_SPECIES
Na 22.989 Na.rel-pbe-spn-kjpaw_psl.1.0.0.UPF
Bi 208.980 Bi.rel-pbe-dn-kjpaw_psl.1.0.0.UPF
ATOMIC_POSITIONS crystal
Bi 0.3333 0.6667 0.2500
Na 0.0000 0.0000 0.2500
Na 0.3333 0.6667 0.5827
K_POINTS automatic
06 06 06 0 0 0
Thanks,
Abhinav
Hello\
Please include pwscf output file and some info: amount of RAM of your mashine and amount of virtual memory.
Also try:
ulimit -a
and if the stack size is too small
ulimit -s unlimited
Sir,
I have changed stack size to unlimited.
configuration of my Pc is:
Ram 4gb and I7 processor.
I have attached the output file.
Hi,
your input file is running well in my old pc. my system configuration is less than yours. i just changed your pseudopotentials(PP) because i don't have that PP u have mentioned in your input file..
@Manoharan: This means that PP wavefcn's are giving a lot contribution to memory requirements. Using different PP will affect memory usage wery much, but that is not always possible.
Yes both are you right and I agree that depending upon pseudopotential computational efforts change . But for this calculation I cant use scalar relativistic potential as I want to perform spin orbit calculations. This file works fine with scalar pseudopotential but I need results using Full relativistic pseudopotential.
Sir Konrad I tried it by changing the stacklimit to unlimited but still getting the same error. Also I low the convergence threshold limit but still no use.
Ok, I'll try to run your input file as soon as my ongoing calculations will be done, but it can take a while.
If everything will go right and your calculation will not crash it will mean that you need to perform this on machine with more RAM, and this sadly cannot be omitted.
Hey Abhinav, could You please send me YOUR'S pp files?
To be very accurate, why it is not working. Or working...
- Badges
- Science topic
- Similar topics
- Physics
- Condensed Matter Physics