Comparing profiling for two full chemistry runs
UM10.2 with `high' optimisation
| Routines |
| UM_SHELL (1,120s) |
| U_MODEL_4A (1,118s) |
| ATM_STEP_4A* (719s) |
UKCA_MAIN1 (360s) |
| ATMOS _PHYS- ICS1 (96s) |
ATMOS _PHYS- ICS2 (88s) |
EG_ SL_ HELM- HOLTZ (40s) |
TR_ SET_ PHYS _4A* (67s) |
EG_CORRECT _TRACERS _UKCA (84s)
|
SL_ TRAC- ER1_ 4A (122s) |
EG_ SL_ MOI- STURE (17s) |
EG_SL_ FULL_WIND (28s) |
STASH (57s) |
See profiling for UKCA_ MAIN1 |
|
Profile for ATMOS_ PHYSICS1
|
See profile for ATMOS_ PHYSICS2 and EG_SL_ HELMHOLTZ
|
EG_SL_WIND_U, EG_SL_WIND_V & EG_SL_WIND_W (4 + 4 + 7 = 15s)
|
STWORK (57s) |
| Itself (25s) |
EG_INTERPOLATION _ETA (92s) |
DEP- ARTURE_ POINT _ETA (14s) |
PP_ HEAD (10s) |
EXP- PXI (17s, itself) |
| EG_ CUBIC_ LAG- RANGE (34s, itself) |
MONO_ ENFORCE (11s, itself) |
Itself (38s) |
Itself (3s) |
*should also link to SWAP_BOUNDS, like many other
returns.
UM10.2 with `safe' optimisation
| Routines |
| UM_SHELL (1,141s) |
| U_MODEL_4A (1,138s) |
| ATM_STEP_4A* (734s) |
UKCA_MAIN1 (363s) |
| ATMOS _PHYS- ICS1 (96s) |
ATMOS _PHYS- ICS2 (88s) |
EG_ SL_ HELM- HOLTZ (41s) |
TR_ SET_ PHYS _4A* (66s) |
EG_CORRECT _TRACERS _UKCA (84s)
|
SL_ TRAC- ER1_ 4A (121s) |
EG_ SL_ MOI- STURE (17s) |
EG_SL_ FULL_WIND (28s) |
STASH (57s) |
See profiling for UKCA_ MAIN1 |
|
Profile for ATMOS_ PHYSICS1
|
See profile for ATMOS_ PHYSICS2 and EG_SL_ HELMHOLTZ
|
EG_SL_WIND_U, EG_SL_WIND_V & EG_SL_WIND_W (4 + 4 + 7 = 15s)
|
STWORK (57s) |
| Itself (s) |
EG_INTERPOLATION _ETA (91s) |
DEP- ARTURE_ POINT _ETA (15s) |
PP_ HEAD (25s) |
EXP- PXI (17s, itself) |
| EG_ CUBIC_ LAG- RANGE (34s, itself) |
MONO_ ENFORCE (11s, itself) |
Itself (s) |
Itself (s) |
*should also link to SWAP_BOUNDS, like many other
returns.
UM10.4 with 'safe' optimisation
| Routines |
| UM_SHELL (1,565s) |
| U_MODEL_4A (1,562s) |
| ATM_STEP_4A* (1,473s) |
| ATMOS _PHYS- ICS1 (196s) |
ATMOS _PHYS- ICS2 (101s) |
EG_ SL_ HELM- HOLTZ (65s) |
TR_ SET_ PHYS _4A* (65s) |
EG_CORRECT _TRACERS _PRIESTLEY (104s)
|
SL_ TRAC- ER1_ 4A (115s) |
EG_ SL_ MOI- STURE (16s) |
EG_SL_ FULL_WIND (33s) |
⇓ |
UKCA_MAIN1 (408s) |
|
Profile for ATMOS_ PHYSICS1
|
Profile for ATMOS_ PHYSICS2 and EG_SL_ HELMHOLTZ
|
EG_SL_WIND_U, EG_SL_WIND_V & EG_SL_WIND_W (4 + 4 + 11 = 19s)
|
STASH (92s) |
See profiling for UKCA_ MAIN1 |
| Itself (33s) |
EG_INTERPOLATION _ETA (90s) |
DEP- ARTURE_ POINT _ETA (20s) |
STWORK (92s) |
| EG_ CUBIC_ LAG- RANGE (31s, itself) |
MONO_ ENFORCE (11s, itself) |
Itself (39s) |
Itself (3s) |
PP_ HEAD (44s) |
EXP- PXI (31s, itself) |
*should also link to SWAP_BOUNDS, like many other
returns.
The extra ~100s in ATMOS_PHYSICS1 looks to be coming from COSP_MAIN.
