[MITgcm-devel] Beaufort experiment on mac os x

[Menemenlis, Dimitris (3248)]

Jean-Michel, the "testreport -fast -a jmc at mitgcm.org"
just completed on my macbook pro.  There was no core meltdown,
but it was close, and as Martin says it was "slow".
I ran two back-to-back testreports.  The first testreport used
latest MITgcm code.  The second incorporated Martin's suggested
changes, i.e., rearrangement of loop in ini_masks_etc:

diff -r1.49 ini_masks_etc.F
119c119,123
<           DO k=Nr,1,-1
---
>          ENDDO
>         ENDDO
>         DO k=Nr,1,-1
>          DO j=1-Oly,sNy+Oly
>           DO i=1-Olx,sNx+Olx

and addition of gmredi_calc_psi_bvp.F in darwin_ia32_gfortran

diff -r1.13 darwin_ia32_gfortran
55c55
<     NOOPTFILES='gad_c4_adv_x.F gad_u3_adv_x.F'
---
>     NOOPTFILES='gmredi_calc_psi_bvp.F'

Both test results were identical and results have been emailed to mitgcm.org.
So from my darwin_ia32_gfortran, older-mac perspective, I'd say go for it.
(I did not do the "-adm" tests,)

Cheers

Dimitris Menemenlis

On Mar 16, 2012, at 6:02 AM, Martin Losch wrote:

>> This beeing said, I still wonder if we might not get other "k-loop insde"
>> optimisation issues in other pieces of the code where it will not
>> create a floating-exception but just wrong results.
>> This was the motivation for this question:
>>> Do we have a full "testreport -fast" output with these flags ?
>>> and if yes, could it be sent to mitgcm.org ("-a jmc at mitgcm.org") ?
>>> I think this could be useful.
> Underway on my laptop, but "-fast" means "slow" (compilation (o:), currently at seaice_obcs, but no problems so far.