Best Practise when loading datasets for graphs

Hello,

We are exploring the most efficient ways to retrieve datasets from our server to display as multiple graphs at once in Perspective.

Since there are two options, tag history binding or Scripting:

Binding:
Does anyone know what settings require the least amount of resources when using tag history binding? f.x. is the query mode pointcount set to 60 less intensive than periodic set to one minute? or is the aggregation mode Average much more intensive than LastValue? Does the binding use a dedicated or shared thread?

Scripting:
Can scripting make it less intensive by f.x. loading the dataset on startup and make a script that adds one row for every new value that comes in? Does it make sense to dedicated a specific thread to this operation?

Is there any significant difference in how we retrieve the data for ApexCharts vs Ignition graphs?