Write another benchmark to measure worst-case block verification latency #3790
Labels
A-benchmarks
Area: Benchmarks
I-dos
Problems and improvements with respect to Denial-of-Service.
I-performance
Problems and improvements with respect to performance
NU1-sapling
Network upgrade: Sapling-specific tasks
NU2-blossom-adjacent
Issues that aren't directly part of the Blossom upgrade, but that we want to complete alongside it.
Milestone
This would be for a 2 MB block full of Sapling spends. (You have to verify both a RedDSA signature and a Groth16 proof for each spend, and the spend descriptions are smaller than output descriptions.)
I think it doesn't matter very much whether these are in a single transaction or multiple transactions. The former is slightly denser in terms of verification work per byte.
The text was updated successfully, but these errors were encountered: