# EPF Week 6 Update This week, I have submitted my project proposal and I have also ran the RISC Zero implementation by Grandine. As per Saulius Grigaitis, the following is the criteria for choosing a zkVM to integrate with Grandine client - The zkVM should be able to prove atleast 100K validators for beachon chain STF - It must have continuation, i.e. being able to prove large programs - It must also have precompiles or some other optimization for `sha2` and `bls12_381`. I've been previously researching on OpenVm but its integration is being handled by the OpenVM core team themselves. So based on the criteria I have been conducting my research on finding suitable zkVMs which can be integrated. So far I have found Ziren (formerly known as zkMIPS) and zkWASM as candidates for my integration. Ziren offers a MIPS architecture with proven performance in large scale computations and continuation support. zkWASM also has a matured ecosystem (WebAssembly target) with existing cryptographic optimizations and recursive proving abilities. I've sent a mail to the corresponding teams, looking forward to their responses. Based on the successful BLS12-381 integration in RISC Zero, I researched precompile strategies for OpenVM: ### Critical Precompiles for Beacon Chain - BLS signature aggregation and verification - SHA-256 hashing for block roots and state roots - Merkle proof verification for validator balances and duties - SSZ serialization/deserialization optimizations Based on the beacon chain STF structure, I've identified optimal continuation boundaries including slot-level processing for individual block state transitions, epoch transitions for validator set updates and finality processing, and large attestation batches when attestation aggregation exceeds memory limits. The recursive proof structure enables individual slot proofs to aggregate into epoch proofs, which combine for multi-epoch verification, maintaining verification efficiency and unbounded computation. ## Work for next week Hopefully by next week, I will be able to complete the zkVM evaluation matrix and finalize on the technical assessment of Ziren and zkWASM against the mentioned criteria, and start working on integrating them or doing further research. ## Conclusion I'm waiting to begin the practical integration work with performance comparison across all the zkVM candidates as soon as I get the results.