Making AsicBoost irrelevant



Summary:

The email exchange between Timo Hanke and Luke Dashjr discusses the idea of redesigning the Bitcoin block header to include a four-byte hash of the first chunk in the second 64-byte chunk. Hanke suggests that taking the midstate hash function may not be necessary since any hash of the first chunk could suffice, and standard software libraries may have more appropriate hash functions. Dashjr asks if XOR-ing bytes 64-76 with the first 12 bytes of the SHA2 midstate would work, but Hanke explains that it would only add another step and make finding a collision marginally harder. Hanke also notes that timestamp rolling on-chip hardware could restrict the redesign to 10 bytes instead of 12. Hanke proposes creating a new mining header with the real blockheader as a child header to address these issues.


Updated on: 2023-06-11T05:17:21.874860+00:00