[Carpet] [Carpet-darcs] CarpetEvolutionMask: fix 2D output for CarpetEvolutionMask_test testsuite

[Erik Schnetter]

On Jan 31, 2007, at 04:23:31, Thomas Radke wrote:

> Erik Schnetter wrote:
>>> Anyway, I believe the testsuite is still passing in principle.  
>>> One  only
>>> needs to fix the header information in the 1D/2D ASCII output  
>>> files  and
>>> recreate the norms in CarpetIOScalar format. Would you agree ?  
>>> If  so I'd
>>> create another patch.
>>
>>
>> Yes, I agree.  Thank you!
>
> I pushed the above patch last week. It turns out that since then the
> testsuite is passing on my laptop and on belladonna, not on peyote  
> though:
>
>     gxx.sum.asc: substantial differences
>        significant differences on 1 (out of 6) lines
>        maximum absolute difference in column 3 is 1.01863406598568e-10
>        maximum relative difference in column 3 is 3.10862224105731e-15
>    Failure: 78 files compared, 43 differ, 1 differ significantly
>
> Is it okay to omit the "sum" norm in this testsuite ? One could also
> increase the absolute tolerance value for the comparison but that  
> would
> affect all other norms as well.

Yes, the sum reduction needs a much higher tolerance, and contains  
essentially the same information as the average reduction.  We can  
just switch it off.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from www.keyserver.net.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 186 bytes
Desc: This is a digitally signed message part
Url : /archives/developers/attachments/20070131/93c6bbe5/attachment.pgp