I would like to support MoL's LinearCombination operator, which is also used to perform time integration on accelerators, to pass an additional argument "refinement level". This will help make it possible to run AMR simulations on accelerators (GPUs). This change is essentially a backport from the respective feature branch.
If not using accelerators, this change affects only MoL.
One of the main drivers for this change is to fold the CaKernel branch back into mainline. Do you agree with such experimental changes to MoL?
-erik
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hello all,
One of the main drivers for this change is to fold the CaKernel branch back into mainline. Do you agree with such experimental changes to MoL?
Fine with me. At the same time, I would also introduce a parameter timelevel to the LinearCombination function to be able to manipulate timelevels other than the current one. This is already proposed in the ticket dealing with MoL's use of accumulators to count evolved etc. variables.
Yours, Roland
- -- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://keys.gnupg.net.
On Feb 18, 2014, at 9:55 , Roland Haas roland.haas@physics.gatech.edu wrote:
Signed PGP part Hello all,
One of the main drivers for this change is to fold the CaKernel branch back into mainline. Do you agree with such experimental changes to MoL?
Fine with me. At the same time, I would also introduce a parameter timelevel to the LinearCombination function to be able to manipulate timelevels other than the current one. This is already proposed in the ticket dealing with MoL's use of accumulators to count evolved etc. variables.
Did that. Thanks!
-erik
users@lists.einsteintoolkit.org
-
Erik Schnetter -
Roland Haas