#928: openmp parallelization within Exact broken ------------------------------------+--------------------------------------- Reporter: knarf | Owner: Type: defect | Status: confirmed Priority: minor | Milestone: ET_2012_11 Component: EinsteinToolkit thorn | Version: development version Resolution: | Keywords: ------------------------------------+---------------------------------------
Comment (by knarf):
A single region implies a barrier at the end, doesn't it?
Of course it would be nice to avoid any barrier if possible. Exact was designed such that the metric functions are 'grid-point local'. They should not depend on something non-local like a 'saved variable'. In the current case, would it lead to regression of performance to simply remove 'firstcall' and 'arad' and use the parameter de_Sitter__scale in the computation of 'am'? What is the reason for having 'arad' in the first place?