You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
These two debug tests take a long time to run (~ 1.5 hours): ERP_D_Lm9.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_monthly and SMS_D_Lm13.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_monthly. Tracing the history of these tests: I had introduced a short decStart transient test with isotopes: ERP_D_P15x2.f10_f10_musgs.IHistClm50Bgc.yellowstone_intel.clm-ciso_decStart, then, in clm4_5_16_r255 (which fixed some C isotope things), @ekluzek introduced something like the above tests (though, at the time, one was on yellowstone and one was on cheyenne). This is also when #177 was found - so maybe he lengthened these to test this issue?
I question whether it's really worth having such long debug tests in the test suite. I guess I can be convinced that there is some value in having a single long debug test in the test suite, but at the very least I'd like to remove this long ERP debug test, replacing it with something like: ERP_D_Ld10_P36x2.f10_f10_musgs.IHistClm50Bgc.cheyenne_intel.clm-ciso_decStart (which is much shorter and covers the crossing of the year boundary).
The text was updated successfully, but these errors were encountered:
Actually, I see that we already have something like my suggested ciso_decStart test, which also includes crops: ERP_D_Ld10.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_decStart. So I'd suggest dropping ERP_D_Lm9.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_monthly and just making that existing ciso_decStart test include threading, by adding P36x2.
These two debug tests take a long time to run (~ 1.5 hours):
ERP_D_Lm9.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_monthly
andSMS_D_Lm13.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_monthly
. Tracing the history of these tests: I had introduced a short decStart transient test with isotopes:ERP_D_P15x2.f10_f10_musgs.IHistClm50Bgc.yellowstone_intel.clm-ciso_decStart
, then, inclm4_5_16_r255
(which fixed some C isotope things), @ekluzek introduced something like the above tests (though, at the time, one was on yellowstone and one was on cheyenne). This is also when #177 was found - so maybe he lengthened these to test this issue?I question whether it's really worth having such long debug tests in the test suite. I guess I can be convinced that there is some value in having a single long debug test in the test suite, but at the very least I'd like to remove this long ERP debug test, replacing it with something like:
ERP_D_Ld10_P36x2.f10_f10_musgs.IHistClm50Bgc.cheyenne_intel.clm-ciso_decStart
(which is much shorter and covers the crossing of the year boundary).The text was updated successfully, but these errors were encountered: