I am having a weird problem with the restarts of the simulations that I am trying to perform. I have decided to perform full restarts with 3 consecutive time-steps. After performing a couple of runs I have realized that the output pressure is messed up everywhere but the velocity looks just good. I realized that this problem started after including the full restart lines in the usr.file:
c if (.false.) then
call blank(fnames,size(fnames)*80)
fnames(1) ='rs_refined0.f00003'
fnames(2) ='rs_refined0.f00004'
fnames(3) ='rs_refined0.f00001'
call full_restart(fnames,3) ! replace istep=0,1,..
c endif
c
iostep_full = 3996
call full_restart_save(iostep_full)
The initialization had been performed with standard restarts and no problems were encountered. I attach some snaps of the velocity (first) and pressure (second). Basically I am simulating the flow past a 3D cylinder and I'm showing you a cut in the z plane (flow in x direction)
Anyone had similar issues? Here the log for a random time-step if that might help:
Step 60000, t= 4.1160376E+01, DT= 2.2500000E-04, C= 0.469 8.2672E+04 1.3010E+00
Solving for fluid
60000 Project VELX 1.5122E+00 3.0714E+00 2.0311E+00 38 38
60000 Hmholtz VELX 2 7.9588E-09 1.5700E-02 1.0000E-08
60000 Project VELY 1.5336E+00 3.2171E+00 2.0977E+00 38 38
60000 Hmholtz VELY 2 9.5493E-09 1.5860E-02 1.0000E-08
60000 Project VELZ 3.8352E-01 2.6593E+00 6.9337E+00 38 38
60000 Hmholtz VELZ 2 4.7842E-10 4.6839E-03 1.0000E-08
60000 Project PRES 1.5286E-01 2.4758E-01 1.6196E+00 15 80
60000 U-PRES gmres 20 9.9268E-06 1.8760E-05 1.0000E-05 2.2606E-01 4.5330E-01
60000 Fluid done 4.1160E+01 1.1386E+00
60000 4.1160E+01 Write checkpoint
2 34 4 kfld
FILE:***Nek5000_refined/run/01_startup/rs_refined0.f00002
min/max: -75.000 125.00 -75.000 75.000 0.0000 5.0000
min/max: -1.2435 1.9696 -1.3373 1.4578 -1.7939 1.2186
min/max: -2.0309 0.48202
60000 4.1160E+01 done :: Write checkpoint
file size = 724.E+01MB
avg data-throughput = 921.3MB/s
io-nodes = 2000