Hacker News Reader

Comment by Supranational | Hacker News Reader

Hey All! Cryptophage team here (https://www.cryptophage.com). Feel free to ask any questions about our alternative approach to solving the puzzle that Bernard just cracked!

View on HN

Congratulations!

One questions: Solving this puzzle has any impact on Verifiable Delay Functions security (specially Chia's VDF[1])?

[1] https://github.com/Chia-Network/vdf-competition

Thanks

Hey MrXOR! First off, thanks for posting this link to HN! Second, the VDF used here leverages RSA groups, while the Chia implementation uses Class groups. These two VDFs are compared here: https://crypto.stanford.edu/~dabo/pubs/papers/VDFsurvey.pdf. Both the VDFs are subject to hardware and algorithmic speedups which will not impact the security of their use as long as these are taken into account. If you are interested in learning more about VDFs I would recommend looking at this collaborative project by Ethereum and Filecoin: https://vdfresearch.org/.

Excuse me, Another question:

From [1]:

>> Cryptophage project solved it in 2 months.

What about ASICs (NSA's power and love!)? How long would it take? any estimation? 2 days? :-)

[1] https://www.cryptophage.com

We are currently investigating ASICs, but a rough estimation would be that they could be at least 10x faster, if not more (e.g. 6 days or less)

Hi there. Can you share some pointers to the alternate algorithm that was mentioned?

Thanks and congrats to you as well for reducing the calculation time!

The goal was to extract as much parallelism from a single squaring operation as possible using the available hardware resources. We used a Xilinx FPGA. There will be a paper detailing the exact algorithm in the coming weeks. In the interim, if you are interested in digging into code, you can find some of the primitives for the design here: https://github.com/supranational/primitives

Reply on news.ycombinator.com