When an Analysis is submitted in SENSE, certain resources need to be allocated for the simulation to run. You are, then, asked to choose the number of Cores and the gigabytes of RAM that you want to use. Choosing the correct values is crucial for the smooth completion of a simulation.
What do Cores represent
CPU cores affect the simulation speed of an analysis. In principle, the more cores you choose, the faster the analysis will run , so you will get the results in a shorter time. + add a rule of thumb like that about RAM.
Fieldscale utilizes smart algorithms that enable an almost linear increase in simulation speed, with the increase in core usage.
What does RAM represent
RAM represents the complexity of a design that can be used.
In principle, adding more gigabytes of RAM will not change anything in speed, but will enable SENSE to compute larger and more complex designs.
As a rule of thumb, solving touchscreen of a 3x3 electrode array design doesn't require more than 8 gb of RAM. However, solving a full design of a 7 inch touchscreen (about 30x20 electrode array) might require 32 or even 48 gb of RAM.
It is up to you to determine how much RAM they will allocate, based on the size and complexity of their design.
How are resources allocated
SENSE allocates resources out of a defined pool of Cores and gigabytes of RAM.
Depending on the account settings, those resources can either be shared with other users or be dedicated for one.
This means that once the pool is depleted and all resources are occupied, an analysis will be shown in a "Submitted" status, until resources become available again.
For example, if you have a pool of 96 Cores and decide to use 40 Cores for an analysis, then 56 Cores will remain in the pool. Now, if you try a new Analysis and request more than 56 Cores, this Analysis will be placed in a "Submitted" status until the requested Cores become available. Finally, if during this time you request with a third Analysis less than 56 Cores, this Analysis will start normally in priority over the second analysis, since the requested resources are available.
You can also submit multiple analyses in the queue. The Analysis that will be selected depends both on submission order and required resources. The first submitted analysis is checked first, if there are enough resources, it begins, else the second is checked and so on.