Changes between Version 28 and Version 29 of ticket/1851/General

Timestamp:

2017-02-27T11:21:00+01:00 (7 years ago)

Author:

frrh

Comment:

--

ticket/1851/General

@@ -214 +214 @@
 This does not happen in the non coupled version, so something to do with MEDUSA coupling specific code?
 
-
+We need to bear in mind that lwm is only ever set to true on PE0 and is used to control may of the namelist prints (though not all it seems (e.g. output.namelist.pis). 
+
+It looks like this bug comes from my fix to prevent the expensive global tracer sums - I'm still writing to  numstr based on lwp when I should be restricting this write to pe zero
+which I could do with lwm, although it's not meant for that, or (if narea==1) or (if numstr > 0) !
+These flags are so confusing. 
+