[MITgcm-devel] tests broken for ifort on faulks

[Jean-Michel Campin]

Hi,

On Fri, Mar 10, 2006 at 04:40:03PM +0100, Martin Losch wrote:
> Hi,
> in case no-one noticed, many tests break for ifort on faulks with  
> this error message:
> (PID.TID 0000.0001) INI_PHIREF: Unable to compute reference  
> stratification
> (PID.TID 0000.0001) INI_PHIREF:  with EOS="POLY3" ; set dBdrRef(1:Nr)  
> to zeros

Nothing to do with this error message: I made this changes because 
I need to compute rho(tRef,Sref,z) and the find_rho_scalar was
giving me plenty of error messages (4 x Nr) when using POLY3.
And I checked carefully that it does not break anything.
I really think that we should tell users not to use POLY3 (and 
probably reduce the Nb of test that are using it).

> this has been going on for weeks now. The last successful test of  
> global_ocean.90x40x15, for example was on Feb23 (and then someone  
> with the name jmc change ini_phiref.F where the error message comes  
> from).

The test on faulks with ifort were OK until yesterday (March 9),
and given the experiments that failed this night, likely to be 
related to mnc.

Jean-Michel