proof_pietrzak Iterations

[HAOYUatHZ]

May I ask why in https://github.com/poanetwork/vdf/blob/master/vdf/src/proof_pietrzak.rs#L45-L58

the iteration number has to be even and >=66? I couldn't find evidence for it in:

• https://eprint.iacr.org/2018/627.pdf
• https://eprint.iacr.org/2018/712.pdf

[jio-gl]

Good question.

[HAOYUatHZ]

We've figured it out.

---

the iteration number has to be even

https://eprint.iacr.org/2018/627.pdf 's idea is to keep halving, so the teration number has to be even

---

the iteration number has to be >=66

• delta is 8 by defalut.
  • https://github.com/poanetwork/vdf/blob/c740ef3b9a4b4663b672ada93bb1c9ba0c80a965/vdf/src/proof_pietrzak.rs#L169
• in generate_proof, when calculate_final_t, the last delta elements will be removed from the !Vec[.... 64, 32, 16, 8, 4, 2, 1], so iterations > 64
  • https://github.com/poanetwork/vdf/blob/c740ef3b9a4b4663b672ada93bb1c9ba0c80a965/vdf/src/proof_pietrzak.rs#L242

---

about delta

this is related to an optimization technique discussed in https://eprint.iacr.org/2018/627.pdf. See section 6.2 Efficiency of the VDF's A minor efficiency improvement