On 12/12/2011 11:35 AM, Frank Loeffler wrote:
On Mon, Dec 12, 2011 at 10:32:55AM -0500, Yosef Zlochower wrote:
-> Internal error: extrapolation in time. variable=LAZEV::dotalptime=118.90625 times=[118.9375]
I am unsure how the times could have gotten out of sync at such a late time in the simulation without getting out of sync earlier. I also do not understand why a time interpolation should have been attempted in the first place. I don't sync or attempt to interpolate this GF.
Do you regrid at that time? It might happen that in order to regrid at that particular time Carpet would need to interpolate in time, and it didn't do before because levels might have been lined up at the time of regridding.
Frank
This is certainly possible, but I thought if you don't sync then no prolongation occurs. Doesn't this mean that 1 timelevel GFs are not regridded (i.e. immediately after regridding, they don't contain valid data and must be reconstructed)?