Is it possible to benchmark agaist blake2 etc. but where they have the same number of rounds, testing both for reducing blake2 and also increasing blake3? Also,

Is it possible to benchmark agaist blake2 etc. but where they have the same number of rounds, testing both for reducing blake2 and also increasing blake3? Also, in that vein, offering the version with more rounds could win over the "paranoid" for mostly being a faster Blake2 thanks to SIMD and extra features thanks to the Merkle tree?