[Developers] checkpointing MoL variables

[Thomas Radke]

Ian Hawke wrote:
> On Mon, 2006-05-15 at 10:51 +0200, Thomas Radke wrote:
> 
> 
>>And the scalar groups MoL_Original_Time and ErrorScalars ? They don't 
>>require much data space in a checkpoint file but their metadata could be 
>>saved.
> 
> 
> I believe that ErrorScalars definitely need checkpointing; I can't see
> anything else that requires checkpointing though.

Okay, I've tagged all variables except the ErrorScalars with 
Checkpoint="no". Tested successfully with one of WaveToyMoL's parfiles 
and BSSN_MoL's Carpet benchmark parfile - the recovered run gives the 
same results as the checkpointed one. So I've committed the change to CVS.

Please report any problems you encounter on this mailing list.

-- 
Cheers, Thomas.