Input files:
setup
NFS=1
output_node_0=0
output_filename=/tmp/slab
derivative_type=2
field_type=1
num_fieldlines=10
NZ=4194304
NX=1
NY=1
nits=4000
lenx=10000
leny=10000
lenz=10000
dt=0.000025
racc=1e-9
Det_t=50.0e0
dbob=0.1
b0=1.0
num_integers=6
num_doubles=4
integer_values
1.0 \\
0.5 \\
0.1 \\
1.666666667 \\
double_values
0 \\
-3897694 \\
-4278780 \\
-200 \\
1 \\
1048576 \\
data
1 5374.35055 5049.64769 9430.07648
10.00000 0.00000 0.00000
2 6708.08852 8934.84235 8485.36849
10.00000 0.00000 0.00000
3 6705.08146 3522.59099 1262.11137
10.00000 0.00000 0.00000
4 6304.60560 7894.47546 6953.63700
10.00000 0.00000 0.00000
5 9241.32407 2839.61296 5638.50522
10.00000 0.00000 0.00000
6 4464.51187 3230.59261 2306.51736
10.00000 0.00000 0.00000
7 1873.51763 2888.45748 9100.62611
10.00000 0.00000 0.00000
8 6482.95939 4790.61276 1300.24582
10.00000 0.00000 0.00000
9 7030.68912 6030.61140 2445.67528
10.00000 0.00000 0.00000
10 7868.83354 5828.88246 4226.56298 10.00000
0.00000 0.00000
Sample of output files:
slab1
uncompressed format (598
KB)
winzip format (276
KB)
gzip format (274 KB)
slab2
uncompressed format (598
KB)
winzip format (276
KB)
gzip format (274 KB)
Accuracy analysis:
Plot between log10 of relative error and t (unit of gyro period)
dbob = 0.0
dbob = 0.001 dbob = 0.01 dbob = 0.1 dbob = 1.0 plot between the relative error at t=1000 gyroperiods and dbob .For each value of dbob, a plot of error is been made for racc = 10-6 and 10-9. Surprisingly, the accuracy in this case is more than tests 1 and 2 for the same value of racc.
For dbob = 0, the accuracy is equivalent with test 1 (constant magnetic field), and when we increase dbob (fluctuations) there is a decrease in the relative error.
The last plot shows the accuracy at the same instant of time for different values of dbob.
