[petsc-dev] Differences between injection and interpolation creation

Barry Smith bsmith at mcs.anl.gov
Tue Jan 13 10:06:45 CST 2015


  I suggest you change the DM to return a matrix in this case and submit a pull request https://bitbucket.org/petsc/petsc/wiki/pull-request-instructions-git then we can get your suggested change into PETSc.

   Thanks

   Barry

> On Jan 13, 2015, at 9:45 AM, Lawrence Mitchell <lawrence.mitchell at imperial.ac.uk> wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi all,
> 
> is there a strong reason why DMCreateInjection builds a VecScatter
> rather than a Mat (unlike DMCreateInterpolation)?
> This seems to rule out creating an injection routine that does coarse
> cell averaging (for, e.g. FV problems).  When used in SNESFAS, the
> scatter from DMCreateInjection immediately gets wrapped up in a Mat
> anyway.
> 
> Cheers,
> 
> Lawrence
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with undefined - http://www.enigmail.net/
> 
> iQEcBAEBAgAGBQJUtT2jAAoJECOc1kQ8PEYvZRsH+wd0xA3Axoojyr65Osrhc3vU
> L/nbVdSRTdFy0wXVo7w6PLhIKpaJ6QE86hZUMHl9EW8JQXLwGn+S7vARS50ElVLL
> I3Ct/1z0RsTujAPxNWwqBym1pTcHShbXZkkPK4x5lkgDsm59XI4dOEGdA4GooBL2
> zLfUQ2hG85dJZJ9G0frm+63wx+Al8qwC4vtWrrUzB04of0iWp8KNYZuQN9aJQ+6R
> LOu4EWGGDB/I+VsaOUl8zrqn3zyh+KQIDpBBdeo4aJ/Tih6QzX17F1OYKA22So0o
> XnBx0ryOMadEB3K6q4g0TS3ItDG0/CKCqmGpOAenfnajCDKrByghZlmxyww52nU=
> =VvW5
> -----END PGP SIGNATURE-----




More information about the petsc-dev mailing list