From 69ee1c8e8b75c325bafb62073ed312298ff02e85 Mon Sep 17 00:00:00 2001 From: Philippe Blain Date: Tue, 16 Aug 2022 08:05:27 -0400 Subject: [PATCH] doc: VP solver is validated with OpenMP When the implicit VP solver was added in f7fd063 (dynamics: add implicit VP solver (#491), 2020-09-22), it had not yet been tested with OpenMP enabled. The OpenMP implementation was carefully reviewed and then fixed in d1e972a (Update OMP (#680), 2022-02-18), which lead to all runs of the 'decomp' suite completing and all restart tests passing. The 'bfbcomp' tests are still failing, but this is due to the code not using the CICE global sum implementation correctly, which will be fixed in the next commits. Update the documentation accordingly. --- doc/source/science_guide/sg_dynamics.rst | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/doc/source/science_guide/sg_dynamics.rst b/doc/source/science_guide/sg_dynamics.rst index 241fa05fe..e6b918538 100644 --- a/doc/source/science_guide/sg_dynamics.rst +++ b/doc/source/science_guide/sg_dynamics.rst @@ -92,8 +92,7 @@ dynamics into CICE is described in detail in The VP solver implementation mostly follows :cite:`Lemieux08`, with FGMRES :cite:`Saad93` as the linear solver and GMRES as the preconditioner. -Note that the VP solver has not yet been tested on the ``tx1`` grid or with -threading enabled. +Note that the VP solver has not yet been tested on the ``tx1`` grid. The EVP, rEVP, EAP and VP approaches are all available with the B grid. However, at the moment, only the EVP and rEVP schemes are possible with the C grid.