?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
(parent: 1261252) | patch
Avoid GPU data race also with OpenCL
commit3fed39606056516812d4ba0e7ad25ae9a8305608
authorMark Abraham <mark.j.abraham@gmail.com>
Mon, 29 Jun 2015 22:32:52 +0000 (30 00:32 +0200)
committerGerrit Code Review <gerrit@gerrit.gromacs.org>
Tue, 30 Jun 2015 10:37:57 +0000 (30 12:37 +0200)
treeefe5838ad865b7013897b7a6fc20da1b365bd44btree | snapshot (tar.gz zip)
parent1261252a63eeac83ad4afbfb60a0de1e9218a066commit | diff
Avoid GPU data race also with OpenCL

Implements the same change to non-local stream synchronization as now
used for CUDA.

Fixes #1756

Change-Id: I720edc0951f97dcff0bd477084fff45a149f01d9
src/gromacs/mdlib/nbnxn_ocl/nbnxn_ocl.cpp diff | blob | blame | history
src/gromacs/mdlib/nbnxn_ocl/nbnxn_ocl_data_mgmt.cpp diff | blob | blame | history
src/gromacs/mdlib/nbnxn_ocl/nbnxn_ocl_types.h diff | blob | blame | history