Benchmarking and unavoidable variance

The cyles/gas the zkVM consumes for an identical blob will vary based on _Celestia's Original Data Square (ODS) size_

- ODS size is variable, thus the _row proofs_ (NMT proof) are going to have different number of layers in the tree - likely order of one or two layers(?). So _each row proof opening_ will have an extra namespace & hash check.
- If the ODS is _mostly full of our blob_ we will will need to include
- (what else...?)

No files, tests, or entry points are named. Start by clarifying how benchmark results should account for Celestia's ODS size, row-proof layers, and namespace and hash checks; the issue does not define a completion criterion.

Written by the indexing model from the issue text.