Skip to content

Commit

Permalink
Enable physics load balancing for CAM-SE by default
Browse files Browse the repository at this point in the history
Recent ACME experiments across all target platforms, mesh sizes,
and process and thread counts indicate that performance
is improved (or, at least, not degraded) by using physics load
balancing option 2 with the point-to-point implementation of
the load balancing communication operator (phys_alltoall == 1).
This changes the defaults to these option settings. Other options,
including the original defaults, can be set in user_nl_cam.

This is bit-for-bit compared to not using physics load balancing
if the BFBFLAG in env_run.xml is TRUE, which is the current ACME
default.

[BFB]
  • Loading branch information
Patrick Worley committed Apr 11, 2017
1 parent 72201fc commit 58cf132
Show file tree
Hide file tree
Showing 2 changed files with 2 additions and 2 deletions.
Original file line number Diff line number Diff line change
Expand Up @@ -1245,7 +1245,7 @@
<iopfile dyn="eul">atm/cam/scam/iop/ARM95_4scam.nc</iopfile>

<!-- physics loadbalance -->
<phys_loadbalance dyn="se"> 0 </phys_loadbalance>
<phys_loadbalance dyn="se"> 2 </phys_loadbalance>
<phys_loadbalance dyn="eul"> 3 </phys_loadbalance>

<!-- Control output to history file(s) -->
Expand Down
2 changes: 1 addition & 1 deletion components/cam/src/physics/cam/phys_grid.F90
Original file line number Diff line number Diff line change
Expand Up @@ -301,7 +301,7 @@ module phys_grid
integer, private, parameter :: modmin_alltoall = 11
integer, private, parameter :: modmax_alltoall = 13
# endif
integer, private, parameter :: def_alltoall = -1 ! default
integer, private, parameter :: def_alltoall = 1 ! default
integer, private :: phys_alltoall = def_alltoall

contains
Expand Down

0 comments on commit 58cf132

Please sign in to comment.