Revision history [back]

We want to run hundreds of thousands of E+ simulations for a parametric study. We have one supercomputer with 6000 cores. But it uses OpenStack architecture, which is different than Amazon AWS. Out IT team says that they won't be able to create a OS-Server on the supercomputer since the framework with Chef, VirtualBox and Vagrant will conflict with our regular operations.

I a thinking of an alternative solution: 1. Install the OS Server on my local computer to generate the .idf files with OS Server's sampling feature 2. Copy .idf files to the supercomputer and run the massive E+ simulations there.

Do you think this is feasible? Or can you recommend better solutions? Thanks,

We want to run hundreds of thousands of E+ simulations for a parametric study. We have one supercomputer with 6000 cores. But it uses OpenStack architecture, which is different than Amazon AWS. Out IT team says that they won't be able to create a OS-Server on the supercomputer since the framework with Chef, VirtualBox and Vagrant will conflict with our regular operations.

I a thinking of an alternative solution: 1. Install the OS Server on my local computer to generate the .idf files with OS Server's sampling feature 2. Copy .idf files to the supercomputer and run the massive E+ simulations there.

Do you think this is feasible? Or can you recommend better solutions? Thanks,

We want to run hundreds of thousands of E+ simulations for a parametric study. We have one supercomputer with 6000 cores. But it uses OpenStack architecture, which is different than Amazon AWS. Out IT team says that they won't be able to create a OS-Server on the supercomputer since the framework with Chef, VirtualBox and Vagrant will conflict with our regular operations.

I a thinking of an alternative solution: 1. Install the OS Server on my local computer to generate the .idf files with OS Server's sampling feature 2. Copy .idf files to the supercomputer and run the massive E+ simulations there.

Do you think this is feasible? Or can you recommend better solutions? Thanks,

We want to run hundreds of thousands of E+ simulations for a parametric study. We have one supercomputer with 6000 cores. But it uses OpenStack architecture, which is different than Amazon AWS. Out Our IT team says that they won't be able to create a OS-Server on the supercomputer since the framework with Chef, VirtualBox and Vagrant will conflict with our regular operations.

I a am thinking of an alternative solution: 1. Install the OS Server on my local computer to generate the .idf files with OS Server's sampling feature 2. Copy .idf files to the supercomputer and run the massive E+ simulations there.

Do you think this is feasible? Or can you recommend better solutions? Thanks,

We want to run hundreds of thousands of E+ simulations for a parametric study. We have one supercomputer with 6000 cores. But it uses OpenStack architecture, which is different than Amazon AWS. Our IT team says that they won't be able to create a OS-Server on the supercomputer since the framework with Chef, VirtualBox and Vagrant will conflict with our regular operations.

I am thinking of an alternative solution: 1. solution:

1. Install the OS Server on my local computer to generate the .idf files with OS Server's sampling feature 2. feature
2. Copy .idf files to the supercomputer and run the massive E+ simulations there.

Do you think this is feasible? Or can you recommend better solutions? Thanks,solutions?

We want to run hundreds of thousands of E+ simulations for a parametric study. We have one supercomputer with 6000 cores. But it uses OpenStack architecture, which is different than Amazon AWS. Our IT team says that they won't be able to create a OS-Server on the supercomputer since the framework with Chef, VirtualBox and Vagrant will conflict with our regular operations.

I am thinking of an alternative solution:

1. Install the OS Server on my local computer to generate the .idf files with OS Server's sampling feature
2. Copy .idf files to the supercomputer and run the massive E+ simulations there.

Do you think this is feasible? Or can you recommend better solutions?