Minor refactoring of CUDA non-bonded kernels
commit07972831f869ff0eb81c8900afbe56c756ce7330
authorSzilárd Páll <pall.szilard@gmail.com>
Wed, 13 Sep 2017 18:37:39 +0000 (13 20:37 +0200)
committerBerk Hess <hess@kth.se>
Mon, 18 Sep 2017 15:53:00 +0000 (18 17:53 +0200)
tree31ebb0c1e1ad921fe34eee3f40d910010d5a8d04
parente02333909decfde2012b81ccc6a9d6bf4f664401
Minor refactoring of CUDA non-bonded kernels

Refactored the dynamic shared memory pointer assignment for improved
robustness. Also reverted some unnecessary leftover changes introduced
with the Volta-enabling and dynamic pruning that resulted in differences
between the legacy and normal CUDA kernels with the goal of improving
maintainability.

Change-Id: I4ce58f3e1f963935d7e3832ed0e537c94d81a632
src/gromacs/mdlib/nbnxn_cuda/nbnxn_cuda_kernel.cuh
src/gromacs/mdlib/nbnxn_cuda/nbnxn_cuda_kernel_fermi.cuh
src/gromacs/mdlib/nbnxn_cuda/nbnxn_cuda_kernel_pruneonly.cuh