# ACCESS-CM2 warm restarts **URL:** \ **Category:** Coupled Model\ **Created:** [1 May 2023 04:20 UTC](https://forum.access-hive.org.au/t/access-cm2-warm-restarts/717 "2023-05-01T04:20:47Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![sebmckenna](https://avatars.discourse-cdn.com/v4/letter/s/3d9bf3/32.png) [@sebmckenna](https://forum.access-hive.org.au/u/sebmckenna)\ **Post date:** [1 May 2023 04:20 UTC](https://forum.access-hive.org.au/t/access-cm2-warm-restarts/717/1 "2023-05-01T04:20:47Z")
Hi all, I am wanting to know what to do if I want to restart an ACCESS-CM2 simulation where it left off? Can you use warm restart to do this? For example you have a run of 10 years beginning in 0951, and I want to restart the model directly following the end can keep generating files the same way (1 year/file). What will the _model basis time_ be? What will the _restart date_ be Restart id/directory will be for the suite that was run for this simulaltion right? Thanks in advance, and ---
**Author:** ![ars599](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.access-hive.org.au/ars599/32/821_2.png) [@ars599](https://forum.access-hive.org.au/u/ars599)\ **Post date:** [3 May 2023 00:38 UTC](https://forum.access-hive.org.au/t/access-cm2-warm-restarts/717/2 "2023-05-03T00:38:53Z")
If the job was successful finished at the year 0960 (ten year run) and you want to extend to run another 10 years, then you can try two steps: 1. change 10 years to 20 years under your suite ID: rose-suite.conf:RUNLEN=‘P10Y’ change to rose-suite.conf:RUNLEN=‘P20Y’ 2. restart the job by using this commend: rose suite-run --restart This is the most easy case. But if you want to start a new job based on the finished 10 years job then you will need to again two steps: 1. same under your rose-suite.conf file: RUNLEN=‘P10Y’ WARM\_RESTART=true WARM\_RESTART\_DATE=‘09600101’ WARM\_RESTART\_DIR=‘where your previous 10 years job location is’ WARM\_RESTART\_NRUN=true WARM\_RESTART\_RUNID=‘your previous 10 year job ID’ 2. rose suite-run --new By the way some of my lazy alias: alias rcgui='rose suite-gcontrol ’ # turn on the RC GUI alias rcnew='rose suite-run --new ’ # start a new job alias rcrd='rose suite-run --reload ’ # if you modified anything, you need to reloaded/update to the job alias rcrg='rose suite-run --reload --no-gcontrol ’ # same as previous one but dont show the GUI alias rcst='rose suite-run --restart ’ # if the job stop not due to crashed, restart your job ---
**Author:** ![sebmckenna](https://avatars.discourse-cdn.com/v4/letter/s/3d9bf3/32.png) [@sebmckenna](https://forum.access-hive.org.au/u/sebmckenna)\ **Post date:** [3 May 2023 00:47 UTC](https://forum.access-hive.org.au/t/access-cm2-warm-restarts/717/3 "2023-05-03T00:47:56Z")
OK thankyou so much for that! I will try the first option as I am wanting to just extend the run from where I left off after testing the experiment is working as expected. Before I use `rose suite-run --restart` do i need to use the `--reload` option that you have written down below in the alias commands? If I encounter any error with this I will update. Do you know if there is any difference with how the model works when restarting from the different methods? ---
**Author:** ![ars599](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.access-hive.org.au/ars599/32/821_2.png) [@ars599](https://forum.access-hive.org.au/u/ars599)\ **Post date:** [3 May 2023 02:47 UTC](https://forum.access-hive.org.au/t/access-cm2-warm-restarts/717/4 "2023-05-03T02:47:24Z")
No no no please use “rose suite-run --restart” Those reload commands apply if the job is running, and in case you change the stop date or modify the namelist, then you will need to use the reload option. This will ensure that at the next time step, the job will incorporate the changes you have made. However, this is not applicable to your case. Those two methods should got identical results. By the way, Shayne McGrager asked me to use GODAS data, and I found it in your folder. For the sake of my lazinessssssssss 🤣, can I just use your data instead of downloading it again? ---
**Author:** ![sebmckenna](https://avatars.discourse-cdn.com/v4/letter/s/3d9bf3/32.png) [@sebmckenna](https://forum.access-hive.org.au/u/sebmckenna)\ **Post date:** [3 May 2023 03:31 UTC](https://forum.access-hive.org.au/t/access-cm2-warm-restarts/717/5 "2023-05-03T03:31:27Z")
Thanks so much. I modified the run length and just used the --restart option. It seems to be running with no issues now. Thanks again Yeah sure please use them if they are useful email or pm me if you need permissions changed. The variables for pottmp, ucur, vcur, dzdt, thflux should all be in `/g/data/e14/sm2435/GODAS/`