On 03/08/2015 02:23 PM, Erik Schnetter wrote: > On Mar 8, 2015, at 13:59 , Yosef Zlochower <[email protected]> wrote: >> >> Thanks Roland. >> >> I should have clarified this to begin with, I was using a single >> processor. >> >> On 03/08/2015 01:44 PM, Roland Haas wrote: >>> Hello Yosef, >>> >>>> We were looking into using the PITTNullCode thorns with Carpet >>>> rather than PUGH. We can compile and run the code, but get thousands >>>> of errors of the form >>>> WARNING[L1,P0] (AEILocalInterp): >>>> CCTK_InterpLocalUniform(): >>>> interpolation point is either outside the grid, >>>> or inside but too close to the grid boundary! >>>> (this may be caused by a global interpolation with >>>> driver::ghost_size too small) >>>> 0-origin interpolation point number pt=0 of N_interp_points=1704 >>>> interpolation point (x,y)=(-0.30024,-0.640512) >>>> grid x_min(delta_x)x_max = -1.46(0.02)-1.28 >>>> grid y_min(delta_y)y_max = -1.46(0.02)-1.28 >>>> >>>> Here the coordinates are defined as grid arrays and range >>>> from -1-eps to 1 + eps (eps is not small). Somehow when >>>> the actual call to AEILocalInterp is made the coordinates >>>> seems to get truncated. >>> The coordinate ranges that you see are those of the local patch of the >>> grid function that AEILocalInterp is using. They are not the global >>> ones. Since the interpolation point is completely outside of the patch, >>> I am actually surprised CarpetInterp even assigns this point to that >>> process at all. >> >> It seems that somehow the extent of the grid array is lost. >> The grid, as seen by AEILocalInterp is only 9 elements >> long. > > Carpet often has a different view of what coordinates are than PUGH. The > flesh offers a coordinate infrastructure, but this doesn't work for AMR. If > you define a local coordinate system for this grid array, then Carpet may be > ignoring this, and instead assume a coordinate system with origin 0 and delta > 1. >
I'll have to dig into the details, but it looks like the origin and deltas are correct, it's the extent of the arrays that isn't. > -erik > >>>> I understand that this type of interpolation may not >>>> make sense in the context of an AMR run, but is it >>>> supported for a unigrid run? I was using carpetInterp. >>>> CarpetInterp2 explicitly does not support grid array >>>> interpolation. >>> One difference may be if PITTNull makes assumptions (implicit) on how >>> PUGH spreads grid arrays among the processes (the DISTRIB=DEFAULT ones). >>> Carpet and PUGH may well differ in that. >>> >>> Yours, >>> Roland >>> >>> >>> >>> _______________________________________________ >>> Users mailing list >>> [email protected] >>> http://cactuscode.org/mailman/listinfo/users >>> >> >> >> -- >> Dr. Yosef Zlochower >> Center for Computational Relativity and Gravitation >> Associate Professor >> School of Mathematical Sciences >> Rochester Institute of Technology >> 85 Lomb Memorial Drive >> Rochester, NY 14623 >> >> Office:74-2067 >> Phone: +1 585-475-6103 >> >> [email protected] >> >> CONFIDENTIALITY NOTE: The information transmitted, including >> attachments, is intended only for the person(s) or entity to which it >> is addressed and may contain confidential and/or privileged material. >> Any review, retransmission, dissemination or other use of, or taking >> of any action in reliance upon this information by persons or entities >> other than the intended recipient is prohibited. If you received this >> in error, please contact the sender and destroy any copies of this >> information. >> _______________________________________________ >> Users mailing list >> [email protected] >> http://cactuscode.org/mailman/listinfo/users > > -- > Erik Schnetter <[email protected]> > http://www.perimeterinstitute.ca/personal/eschnetter/ > > My email is as private as my paper mail. I therefore support encrypting > and signing email messages. Get my PGP key from https://sks-keyservers.net. > -- Dr. Yosef Zlochower Center for Computational Relativity and Gravitation Associate Professor School of Mathematical Sciences Rochester Institute of Technology 85 Lomb Memorial Drive Rochester, NY 14623 Office:74-2067 Phone: +1 585-475-6103 [email protected] CONFIDENTIALITY NOTE: The information transmitted, including attachments, is intended only for the person(s) or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and destroy any copies of this information. _______________________________________________ Users mailing list [email protected] http://cactuscode.org/mailman/listinfo/users
