Simply run ./run.sh to execute a benchmark for each of the proving systems
- Cairo with
Stoneprover - Noir with
Barretenbergprover - Rust with
sp1 - Rust with
zkllvm
The current benchmarking program is exponential fibonacci, with a default of 20 iterations, giving about 2^20 number of function calls.
Because the shape of the circuit is dependent on input, this is not optimal for building programs.
For this reason workarounds have to be used, as in the generics.nr