Changes between Version 61 and Version 62 of ticket/1851/General
- Timestamp:
- 2017-03-10T13:03:39+01:00 (7 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
ticket/1851/General
v61 v62 494 494 have been able to spot these sorts of bugs, w/o some sort of dummy interface. 495 495 496 496 I've tried knocking up a fix by adding the rnf_b = rnf assignment to sbcmod.F90 with all the other similar TS related shuffling and protecting it with IF (ln_rnf) and protecting the original shuffle in the wrong place as far as coupling is concerned with IF (.NOT.ln_rnf). That works well and we finally have solver.stat agreeing at 2 days. 497 However Tim has an even simpler solution which also seems fine: branches/UKMO/dev_r5518_GO6_package_fix_rnf@7782 498 499 We need to get this reviewed and included asap. Tim has created a separate ticket #1866 since this is a general NEMO issue, not just something peculiar to MO systems. 500 501 Someone (probably I) also needs to come up with a viable solution to the CICE problem. 502 503 With all these hacks in place we can finally see that although NEMO and CICE NRUN v CRUNS agree at 2 days, the CO2 field coming from MEDUSA to be used in coupling is diverging on day 2. 504 But it doesn't affect evolution.... why? MS confirms that in this set up the CO2 field is effectively only a diagnostic, but it will be employed as a prognostic in other runs. 505 506 So we need to get that right! 507 508 Well we already know that MEDUSA diverges in stand alone runs too so it'll be easier to debug in that context probably. 509 510 In fact MEDUSA developers really ought to be best placed to do that. 511 512 I also ought to be updating things to be using the vary latest GO6 and MEDUSA branches since things have moved on a lot even in the ~3 weeks since I started this. 513 514 515 516 517