Hi Anthony, I also ran across this issue recently, and I believe it is caused by the TotalQuantity derived quantity returning an integer 0 instead of 0.0. I just pushed this with another minor change that may have helped as well. If you update your install, or in this case, since you're on kraken, have the admins update the yt module, it should be fine now. Please, let us know if it's not, though. Britton On Thu, Jul 14, 2011 at 12:41 PM, Anthony Harness < anthony.harness@colorado.edu> wrote:

I am trying to use the derived quantity, 'TotalQuantity', to count the number of cells included in my cut_region as follows:

cut = dd.cut_region(["grid['x'] > 0.5"]) #Example cut numberCells = cut.quantities['TotalQuantity']('Ones')[0] #Totals a 'one' for each cell, which gives the total number of cells being used

This works fine when I have plenty of cells to split up between 84 processors (or when there are exactly zero cells in the cut), but when there is a small number of cells in the cut (<~300), I get erroneous (way too large) random values like 1.e18 or -4.2e17. I know there are cells in the cut because I don't get a 'nan' when I calculate a 'WeightedAverageQuantity' value. I have tried to use "cut['Density'].size" to get the number of cells, but I get killed by the OOM Killer. My original data set has 1024^3 cells and I am running on Kraken with 84 processors.

Thanks, Anthony

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org