Speaking in a very general sense, most of the time there is a trade-off:
- Lower in-memory usage, but slower access (because items need to be dynamically processed).
- Faster access, but more memory usage.
This trade off can easily be seen via things like math lookup tables. You can run the function pow(2, x) every time you need it, or if you know that x will always be a positive integer, you can build a table ahead of time and look it up instead. The former uses less memory but is much slower, the latter uses more memory and is much more limited, but is a lot faster.
The conclusion, therefore, is to draw little connection between ram consumption and performance.
