Parallelization of zk-Proving Process for Enhanced Performance in Scroll Prover

[drnick30]

The current proving process in scroll-prover appears to handle transactions sequentially, which may be a bottleneck for more complex proofs. I suggest exploring parallelization techniques, particularly by offloading certain proof-generation steps, such as FFTs or polynomial commitments, to multiple cores or even GPUs. Implementing this could significantly reduce proving time for batch transactions. I’d be happy to assist with initial benchmarking or testing parallelization methods to assess potential performance gains.