Performance of Standard Configurations
Global Atmosphere 7.0/Global Land 7.0
Suite id | Description | UM Version | Date Run | Platform | Decomposition* | OpenMP Threads | Length of run | Dump Frequency | Wallclock | Data output vol | Comment | Cost/model-yr |
---|---|---|---|---|---|---|---|---|---|---|---|---|
u-ab878 | GA7.0 N96 AMIP | 10.3 | 10.02.17 | ARCHER | 16 x 12 | 2 | 30 days | 10 days | 01:07:53 | |||
u-ag689/n216 | GA7.0 N216 AMIP | 10.5 | 22.06.17 | ARCHER | 42 x 36 | 1 | 10 days | 10 days | 00:29:27 | |||
u-ak155 | GC3.1 N216 ORCA025 | 10.6.1 | 07.06.17 | ARCHER | A:42x36 O:24x26 X:12 (6 nodes) | 1 | 30 days | 10 days | 01:24:29 | |||
u-ai862 | GC3.0 N96 ORCA025 | 10.6 | 20.07.17 | ARCHER | A:16x18 O:18x26 X:8 (4 nodes) | 1 | 10 days | 10 days | 00:27:03 | |||
u-aq711 | piControl N216-ORCA025 | 10.7 | 17.10.17 | ARCHER | A:36x24x2 O:30x30 X:8 nodes (16/2) | 2 (atm) | 1 mnth | 10 days | 1:45:00 | 25% variation in run times - expensive because multi-threaded (necessary?) working copy only | ||
u-as037 (r59528) | GC3.1 N96 ORCA1 PI Control CMIP6 | 10.7 | 11.11.17 | ARCHER | A:12x8 O:9x8 X:6 (3 nodes) | 2 | 1m | 10 days | 00:56:59 | 162Gb (History_Data - 2month run) | ||
u-au173 | GA7.0 N96 UM10.9 AMIP Climate Development (1988 - 2008) | 10.9 | 17.01.18 | ARCHER | A:16x24 | 1 | 30 days | 10 days | 00:36:17 | |||
u-av321 | GA7.0 N96 UM10.7 AMIP Climate Development (1988 - 2008) | 10.7 | 23.02.18 | ARCHER | A:16x24 | 2 | 30 days | 10 days | 00:30:47 | |||
u-ba767 | GA7.0 N96 UM11.1 AMIP Climate Development (1988 - 2008) | 11.1 | 31.08.18 | ARCHER | A:14x18 | 2 | 30 days | 10 days | 00:33:05 | Note-average time over 1st 3M | ||
u-ba767 | GA7.0 N96 UM11.1 AMIP Climate Development (1988 - 2008) | 11.1 | 04.09.18 | ARCHER | A:24x24 | 2 | 30 days | 10 days | 00:26:25 | Note-average time over 1st 3M- (00:29:35, 00:24:34, 00:25:06) | ||
u-bd431 | GC3.0 N96 ORCA025 Present Day | 11.2 | 28.11.18 | ARCHER | A: 12x16 O:18x26 X:8 (4 nodes) | 2 (atm) | 1 month | 30 days | 01:11:55 | 39 Gb (History_Data) |
UKESM
Suite id | Description | UM Version | Date Run | Platform | Decomposition* | OpenMP Threads | Length of run | Dump Frequency | Wallclock | Data output vol | Comment | Cost/model-yr |
---|---|---|---|---|---|---|---|---|---|---|---|---|
u-an561 | UKESM0.6-CN PI, new oxidants file, lana, StratTrop | 10.6 | 06.29.17 — 07.17.17 | ARCHER | A:48x24 O:9x12 | 1 | 1yr (12 x 1 mnth cycles) | 10 days | 12 x 52 mins | 111GB (archived (no dumps, pp+compressed netcdf)) | ||
u-q606 | UKESM0.8-CN N216 ORCA025 (some improvements) | 10.7 | 04.10.17 | ARCHER | A:36x24 O:30x30 X:4 nodes (8/2) | 1 | 1 mnth | 10 days | 2:00:00 | working copy only | ||
u-az760_8.3.7-GL | UKESM1 alpha release | 11.0 | 16-09-18 | ARCHER | A:48x24x1 O:9x8 X:3 nodes (6/2) | 1 (atm) | 1 mnth | 30 days | 00:49:00 | ~60GB | working copy only - not for production runs | 190 kAU |
u-bc108 | UKESM1.0-beta historical | 11.1 | ARCHER | A:48x24x1 O:12x8 X:3 nodes (6/2) | 1 (atm) | 1 mnth | 30 days | not yet for production runs | kAU | |||
u-bc409 | UKESM1.0-beta pre-industrial control | 11.1 | 06.11.18 | ARCHER | A:48x24x1 O:12x8 X:3 nodes (6/2) | 1 (atm) | 3 mnth | 3 mnth | 01:53:00 | ~86GB | not yet for production runs cce8.3.7 | 158 kAU |
u-bd204@99131 | UKESM1.0-beta pre-industrial control | 11.1 | 04.12.18 to 19.12.18 | ARCHER | A:48x24x1 O:12x8 X:3 nodes (6/2) | 1 (atm) | 30 years | 3 mnth | 251 hours | 5.6TB | - | 165.7 kAU |
*Decomposition: A = Atmos (EW x NS), O = Ocean (EW x NS), X = XIOS processes (nodes)
Scaling & Cost Plots - ARCHER
GC3.1 N96 ORCA1 u-as037 (r 61036)
This looks at the performance of HadGEM3-GC3.1 N96ORCA1 PI Control for CMIP6 over a wide range of atmosphere decompositions with a fixed number of IO servers (6). The model was run on ARCHER in December, 2017. The ocean decomposition was 9x8 and the XIOS used 6 processes. The atmosphere was run with two threads, except for one run at 36x28 where three threads were used (3,186 processes in the graph below).
The following graphs show the model rate (in model years per day) against cost and total number of processes. Each point in the graph is the result of running the coupled suite for a model month. The cost axis was obtained by rescaling the processors axis in the right-hand graph. The cost of computing the graph was 440 kAU.
![]() | ![]() |
GA7.0 (u-ba767)
Similar calculations are performed for GA7.0. Real costs are used. The performance appears badly affected in the high processor limit.
![]() | ![]() |
UKESM AMIP (u-bh142)
The scaling graph as a function of the number of CPUs and the cost per model year is shown. Orange (1) and blue (2) are the number of OMP threads respectively. Further details, such as the North-South and East-West decomposition are revealed by holding the mouse above a point. Data from ARCHER, with a cost 0.015 kAU per core hour.
Attachments (7)
-
Rplot05.jpeg
(56.5 KB) -
added by willie 3 years ago.
Model Rate vs Cost
-
Rplot06.jpeg
(55.0 KB) -
added by willie 3 years ago.
Model rate vs total processes
- GA7costGraph.png (21.5 KB) - added by dcase 2 years ago.
- GA7cpuGraph.png (20.0 KB) - added by dcase 2 years ago.
- GA7costGraphNew.png (21.5 KB) - added by dcase 2 years ago.
- GA7cpuGraphNew.png (20.0 KB) - added by dcase 2 years ago.
- profileOutput.html (6.6 KB) - added by dcase 2 years ago.
Download all attachments as: .zip