Closed AaronFeickert closed 1 year ago
Should proving complexity be considered important enough to include? They're often given much less weight since they're typically one-off operations whose complexity is irrelevant to verifiers. A one-time increase of a few milliseconds (to a prover) doesn't matter in the same way that a million increases of a few milliseconds (to a verifier) would be.
I think it is important to add something, yes. It can always be put into context, as you pointed out. I would imagine that on some devices proving time would matter, like on a mobile phone when transactions are created.
Description
Updates range proof RFC-0181 to add comparative performance data and improve notation.
Closes #102.
Motivation and Context
An earlier PR removed outdated benchmark data, but it was suggested to include reasonable comparative performance data against Bulletproofs. This data has been added for both proof size and verification complexity.
Some notation was incomplete or incorrect. In particular, sum and index bounds were left implicit in many places, and indexing relating to mask recovery was incorrect. Both of these have been addressed.
How Has This Been Tested?
The changes build and appear to render correctly.