Progress count starts to show negative values from some point of simulations

naotohori commented 4 years ago

The progress information in the standard output can be broken after some point of simulations. Have to check write_progress.F90.

41 # PROGRESS shwon every 1000000 steps 42 #1: step 43 #2: million steps / day (latest cycle) 44 #3: million steps / day (averaged) 45 #4: remaining time / hours ....... 546 500000000 266.501 272.752 44.0 547 501000000 265.655 272.737 43.9 548 502000000 265.814 272.723 43.8 549 503000000 265.727 272.709 43.7 550 504000000 -.402942 -790.854 -15.1 551 505000000 265.833 -797.129 -14.9 552 506000000 266.171 -803.472 -14.8 553 507000000 267.025 -809.876 -14.6 554 508000000 269.043 -816.320 -14.5 ...... 716 670000000 268.011 -51005.8 -0.2
717 671000000 268.206 -71327.8 -0.1 718 672000000 268.159 -118349. -0.1 719 673000000 268.001 -345692. -0.0 720 674000000 268.247 378422. 0.0 721 675000000 268.210 122515. 0.1 722 676000000 268.164 73171.5 0.1 723 677000000 268.001 52197.7 0.1

naotohori commented 4 years ago

References

https://software.intel.com/en-us/fortran-compiler-developer-guide-and-reference-system-clock

https://stackoverflow.com/questions/46637148/do-i-use-fortran-system-clock-correctly/46639558#46639558

naotohori commented 4 years ago

Integer type is changed to L_INT in write_progress.F90.
Fix time_integral_post so that it works when restart.
Use ISO_FORTRAN_ENV

Hori-Lab / TIS

Progress count starts to show negative values from some point of simulations #4