MITgcm / verification_other

Additional verification experiments for MITgcm
MIT License
10 stars 19 forks source link

minor update to reference llc90 ADM output #38

Closed jm-c closed 3 years ago

jm-c commented 3 years ago

Previous new ADM output (from PR #37) were generated on engaging using 16.procs instead of on 32 procs as it was done before. This does affects AD var stats and cost-function results (at machine truncation level). Stick to 32.procs for this AD experiment reference output and update (again) these 4 output_adm*.txt files.