| commit | b9e713b61cd1389b9b32f1d4677f252b90cd0379 | |
| author | Jonathan Vincent <jvincent@nvidia.com> | |
| Mon, 1 Oct 2018 09:59:40 +0000 (1 02:59 -0700) | ||
| committer | Mark Abraham <mark.j.abraham@gmail.com> | |
| Wed, 31 Oct 2018 16:47:43 +0000 (31 17:47 +0100) | ||
| tree | 1cfc4c0a4e09f6c2414d08a89a408037800a8b58 | treesnapshot (tar.gz zip) |
| parent | 631a25cf0f7fc1f61ed49c3f444416480e4c132d | commitdiff |
Add CUDA bonded kernels
CUDA bonded kernels are added for the most common bonded and LJ-14
interactions.
The default auto settings of mdrun offloads these interactions
to the GPU when possible.
Currently these interactions are computed in the local or non-local
nbnxn non-bonded streams. We should consider using a separate stream.
This change uses synchronous transfers. A child change will change
these to asynchronous.
Updated release notes and performance guide.
Fixes #2678
Refs #2675
Change-Id: Ifc6d97854cc7afa8526602942ec3b1712ba45bac
CUDA bonded kernels are added for the most common bonded and LJ-14
interactions.
The default auto settings of mdrun offloads these interactions
to the GPU when possible.
Currently these interactions are computed in the local or non-local
nbnxn non-bonded streams. We should consider using a separate stream.
This change uses synchronous transfers. A child change will change
these to asynchronous.
Updated release notes and performance guide.
Fixes #2678
Refs #2675
Change-Id: Ifc6d97854cc7afa8526602942ec3b1712ba45bac
24 files changed: