Monitor#

Simulation time#

cloudHPC provides an estimate - ETA - of the total simulation's duration. This is provided after the simulation has been executed and it is available in your Output as highlited by the image below. The ETA estimate is available only for the following software:

• FDS [all versions]

Manual estimate#

To manually evaluate the simulation’s duration, you can check how many iterations are pursued and in how much time. This estimation can be done after at least 2 hours of simulation. You can then evaluate how many seconds of simulation are pursued in real-time. Hence, for instance, how many seconds of simulation are calculated in 2 hours. If you want to pursue a simulation of 600 seconds and you valuate that a simulation time of 120 seconds is calculated in 2 hours analysis. Then you would know that your simulation would take approximately 60 seconds of simulation time each hour. Consequently, it would take 10 hours to complete a simulation time of 600 seconds.

120s : 2h = 60s/h 600s : 6s/h = 10 h

Runtime monitor#

The cloud HPC system, offers you the possibility to monitor your analysis in runtime during the simulation and, at the same time, to check the usage of the hardware resources in terms of vCPU and RAM, during the analysis. In order to access the monitor you need to enter the simulation details by clicking the "View" button of the analysis you want to monitor.

Once you enter the simulation view page, the cloudHPC provides you with a number of information and controls of the running analysis that are briefly expained by the following paragraphs.

Simulation output#

The most fundamental communication window is represented by the simulation output. This is a real-time log of any information produced by cloudHPC while your simulation is actually running. In these info you can find:

1. Errors and warnings produced by the cloudHPC
2. A basic plot of the runtime log of your executable
3. Execution information of the cloudHPC at all phases of your analysis

The output is the main window available under the simulation view page as displayed by the following image.

Runtime plots#

As your solver runs, cloudHPC tries to plot the most important output into graphs for users to easily read them. Every CSV file generated will be automatically plot into graph. Beyond that, some software have the capability to produce more advanced outputs such as:

• OPENFOAM: the cloud HPC authomatically generates a chart for each runtime post-processing activated in controlDict. It is important that you set it before launching the analysis with the post-processing you want to detect: residuals, max pressure/velocity, temperatures, flow calculations, etc. In order to have a template of the controlDict file you can check here.

• FDS: they system authomatically convertes into chart every CSV file generated by FDS during the runtime. In order to see the results you need to set up &DEV (devices) and properly assign them variables to monitor.

Runtime logs#

Every simulation software produces a log file which is displayed directly at runtime in the proper section of the web app. It is possible for you to check these logs in two ways:

1. Through the Output section where you can see a brief summary of the simulation status
2. In the Logs section where every log file produced is reported and made accessible. The access remains also when the simulation is terminated as a reference

Remote desktop#

For enabled accounts, it is possible to access the virtal machine where the simulation is actually performed via remote desktop. This feature is available under the "view" page of each simulation and gives you access to all hardware resources of the instance where simulation is performed. With that you can:

• monitor your running process
• check status of the hardware resources
• run your preferred post-processing software such as ParaView, SALOME or Smokeview

The remote desktop is made available for all simulations where status is RUNNING. To access it just enter the "view" page of the simulation and scroll down at the bottom of the page.

Hardware monitor#

Beyond results extracted by the solver, at any time it is possible to monitor the usage of vCPU and RAM are. These graphs are important in order to understand if the computational capacity chosen for your simulation is enough or overestimated.

As a general rule, you can try to always have at least 1000 Mb available in the Free line (red).

SSH Connection#

Advanced users can also connect via SSH to the instances where the simulation is running. This allows you to enter the instance and get full control of the running analysis. You are also given the possibility, for example, to create/edit/cancel files.

Prerequisites#

In order to get SSH access to the running simulations make sure you have correctly saved your public key into your profile such as described in the following image:

In order to retrieve your public key on Linux based systems, you can just run the following command on a terminal:

cat .ssh/id_rsa.pub

Connect via SSH#

Once you have saved your public key to access via SSH to a running instance type the following command:

ssh -X -o "StrictHostKeyChecking no" cloudhpc@SIMULATION_IP_ADDRESS

where SIMULATION_IP_ADDRESS is reported in the "view" page of the running analysis. The same result can also be achieved using the script cloudHPCexec by typing:

cloudHPCexec -ssh SIMULATION_ID