Improve estimate_module_size calculation and doc comment

[webmaster128]

We need some updated estimations on how much memory a module uses in the in-memory cache.

[webmaster128]

This percentage (~100.06%) was derived by comparing the serialised size to the size of the memory reported by massif. As such, is not considering mmaped memory, used in the module's storage.
Better to remove this comment, or update it to account for the size discrepancy due to mmaped storage.
We can also introduce a factor to multiply this module size, to get a better estimation using the old method.

@maurolacy do you know if additional mmaped memory is used here at all? I know modules are archived to disk using kryv so I would expect that the on-disk size is aktually a very good estimation of the in-memomory size.

[webmaster128]

Okay I found #959 (comment) again, which explains things.

What we can do is create a patched CosmWasm 1.2 which exports loupe size and module size to determine a reasonable factor for real-world data.

[maurolacy]

Okay I found #959 (comment) again, which explains things.

What we can do is create a patched CosmWasm 1.2 which exports loupe size and module size to determine a reasonable factor for real-world data.

That's a good idea. Based on the comment's data above, an initial (rough) factor is ~1.5.

Another option is to leave this aside, and just use the artefact size (disregard the store size, that is) as a proxy for the entire thing's size. Mentioning this for completeness. It has the benefit that is simple, it works, and it's what we already have.