- 09 Mar 2023
- 4 Minutes to read
How to configure a cloud account
- Updated on 09 Mar 2023
- 4 Minutes to read
On the Flood Modeller interface, select the 'Simulations' tab and click on 'Run Flood Cloud'.
To configure your account, click 'Settings' on the Flood Cloud toolbar which will display the 'Cloud setup' window and provides five tabs.
1) Login tab
Enter your login details and select 'Retrieve'.
2) Resources tab
Slide the on/off switch to the 'On' position. The 'number of machines to use' box allows you to set a maximum value only. This is limited to a range of 1 to 50 machines. It is recommended to leave this setting at the default value of 50. The system will only utilise the number of machines it needs to run a batch (rather than always starting all 50).
Note that, after logging in, the settings on the 'Resources' tab will display the last used values for the specified account (which may have been set by a different user as accounts can be shared by multiple users).
3) Options tab
The 'Options' tab provides an option to automatically download results. This is enabled by default allowing model results to be downloaded immediately as each simulation in your batch completes. If you untick this option, then all results will be stored in your assigned cloud storage space (limited to 50GB) and you can then manually select which completed simulations you then download results from (using the 'Download' button in the toolbar of the Flood Cloud interface).
Associated to the results download is a default folder location on your system. This provides a backup location for results to be downloaded to if Flood Modeller cannot identify a results folder specification within your selected simulation files.
The job identifier text is also defined on the 'Options' tab. It is recommended that for each batch you run in Flood Cloud you enter a unique name in the job identifier box. This name is then used later in the run history interface, enabling you to access progress, review server time used for the job or download results. A check-box is provided to ‘Prompt for identifier when each job is submitted’ – it is recommended to tick this box to ensure each batch has a unique name.
The 'Show solver window when running simulations' checkbox relates to models that include TUFLOW. Some models can pause waiting for user interaction (in most cases these user interactions are bypassed by Flood Cloud). When a pause happens, it can lead to cloud credits being wasted by the simulation. Pauses are usually due to incorrect model setups, therefore the test mode should be utilised to check models locally before they are uploaded to the cloud.
If you experience a problem with a model when testing you can re-run with the 'Show solver window when running simulations' checkbox ticked. The simulation will then be configured to display the native TUFLOW progress window (and Flood Modeller progress window if a linked model), making any prompts for user interactions immediately apparent. Issues with the models can then be addressed prior to starting any cloud simulations (avoiding loss of credits).
4) TUFLOW tab
The ‘TUFLOW’ tab allows you to define TUFLOW's solver precision setting i.e. single or double. Single precision is the default setting. Note that if the precision has been set within your TUFLOW control file, then these settings will override your Flood Cloud interface setting.
Flood Cloud provides three options for packaging up your TUFLOW model files:
- Standard copy: a TUFLOW function that copies all model input files to a new location
- TUFLOW model packager: a TUFLOW function that will analyse your model and locate all required input files
- Folder upload: a Flood Cloud function that copies all data into a specified folder and requires a common folder structure
In addition, if you select one of the packaging methods provided within the TUFLOW software, you also have the option of selecting a TUFLOW version to perform this packaging. You can use the version included with Flood Cloud (the default setting) or use your locally installed version of TUFLOW. Note that the selected version does not define the version of solver that will be used for the subsequent simulation in the cloud.
For the majority of model scenarios, the included version of TUFLOW packaging tools should work and the 'use locally installed' option is only recommended for those rare cases when the included packager has failed (and you don’t want to modify your model files into a single folder setup to utilise the 'Folder upload' option). Note that an active local licence may be required for the packaging of your models using your locally installed version of TUFLOW (this will depend on the version number of your copy of TUFLOW).
Other TUFLOW related Flood Cloud settings presented on the 'TUFLOW' tab are:
- No licence is taken when packaging TUFLOW models. Uncheck this box if you instead wish to utilise a local licence. This may be required if you have selected to use a copy or packager function from your locally installed TUFLOW (instead of the version included within Flood Cloud). Note a local licence must therefore be available when you run your Flood Cloud simulations in this case
- Check 'Run in ‘no console’ mode' to make sure that any simulations on the cloud do not “hang” waiting for input from the “user”
It is recommended you initially keep the default settings for all of these options and only change these settings if you experience any issues.
5) Advanced tab
These settings should be left with all default settings unchanged.