Derecho scaling issue: cesm1.3 #840
jedwards4b
started this conversation in
General
Replies: 3 comments 6 replies
-
Does the ihesp version of CAM use the new infrastructure mods that CAM-SE uses? |
Beta Was this translation helpful? Give feedback.
4 replies
-
Changing the lbal_opt from 2 to 1 made a difference. I'm not sure that updating the ihesp code to use the latest cam is realistic - is it? |
Beta Was this translation helpful? Give feedback.
1 reply
-
I don't suppose anyone has a program that isolates and tests the performance of lbal_opt in cam? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I am getting really poor scaling from an ne120 FC5 using the ihesp version of cam on derecho.
Scaling plot is here. I have traced the issue to phys_grid.F90:transpose_block_to_chunk
There are several options to control this routine including lbal_opt and a CPPDEF MODCM_DP_TRANSPOSE
Can someone suggest what options I should explore first?
Beta Was this translation helpful? Give feedback.
All reactions