These need to be preserved when changing the per-instance problem parameters (for example: In the N-Body simulation: time simulated, where bodies are, the approximation distance). If the items in the slide are not preserved, then the scaling measurements are not valid.
am I right?
Edit: Ah, going forward in the slide deck, I see that scaling down the problem while preserving these properties allows measurements to be done which accurately reflect how a larger problem will behave.
Just to clarify:
These need to be preserved when changing the per-instance problem parameters (for example: In the N-Body simulation: time simulated, where bodies are, the approximation distance). If the items in the slide are not preserved, then the scaling measurements are not valid.
am I right?
Edit: Ah, going forward in the slide deck, I see that scaling down the problem while preserving these properties allows measurements to be done which accurately reflect how a larger problem will behave.
This comment was marked helpful 0 times.