当我并行运算的时候,将CPU的数量翻倍,会将每个CPU上所需的内存减半。我运行例子中的IN.MELT时的确如此。可见并行是成功地。但当我使用自己编写的in.BiTiO时,翻倍CPU数量,则每个CPU的内存几乎没有变化。这意味着我无法计算更大的系统。不知道是不是因为我使用了KSPACE EWALD,少安装了某个PACKAGE或者其他什么原因,求高人指点,万分感激
#Initialization
atom_style full
units metal
dimension 3
newton on on
boundary p p p
read_data BiTiO333
#setting
kspace_style ewald 1.0e-6
pair_style buck/coul/long10
#Potential set A
pair_coeff * * 0 0.3010 0
pair_coeff 1 3 49529.35 0.2223 0
pair_coeff 3 3 24244.50 0.3284 0
pair_coeff 1 1 9547.960.2192 32.00
pair_coeff 1 2 2549.4 0.2989 0
run_style verlet
velocity all create 300.0 539 #assign the initial velcocity at 300K
#start the loop and output
include tempfile1
reset_timestep 0
variable i loop 5 #make 5 loops
label loopa
#Equilibrate at certain temperature
reset_timestep 0
fix npt all npt temp ${mytemp1} ${mytemp1} 0.1 aniso 0.0 0.0 0.1#the damping parmeter 0.05 here specifies that tempreature/pressure shall relax in 0.1 time unit
run 5000
reset_timestep 0
#Obtain cell information
thermo_style custom step temp vol ke pe lx ly lz etotal press
thermo 100
variable lx equal lx
variable ly equal ly
variable lz equal lz
variable vol equal vol
variable press equal press
fix time_average all ave/time 1 500 500 c_thermo_temp v_lx v_ly v_lz v_vol v_press file output${mytemp1}.txt #Take time average of the results
run 500
unfix npt
unfix time_average
write_restart ${mytemp1}.equil
next mytemp1
next i
jump in.BiTiO loopa