SIMD-0052: Add Receipt Root to Bankhash

[anoushk1234]

This SIMD is the first step towards implementing a light/diet client for Solana as described in #10 and
and also completes the implementation of SPV. This would ensure that users don't have to trust their RPCs and can instead directly trust the staked nodes.

[aeyakovenko]

If we want a succinct proof that the majority computed an invalid output we need the receipt to contain the input account version for all the accounts. read and write.

Account version is the number of txs that had this account with a write. Then it would be easy to prove if the majority used a old version, or if the output was incorrectly computed.

[anoushk1234]

If we want a succinct proof that the majority computed an invalid output we need the receipt to contain the input account version for all the accounts. read and write.

Account version is the number of txs that had this account with a write. Then it would be easy to prove if the majority used a old version, or if the output was incorrectly computed.

@aeyakovenko I like the idea of having a succinct proof but as discussed, this should be a separate proposal and change.

[aeyakovenko]

@anoushk1234 what is the performance impact of adding the status code into the current labs merkel tree computation for the bank hash?

[anoushk1234]

@anoushk1234 what is the performance impact of adding the status code into the current labs merkle tree computation for the bank hash?

@aeyakovenko We did some benchmarks, so calculating a root for 1 million receipts using the labs merkle tree takes ~ 344ms - 397ms depending on the machine.

Given that we will see other parts of the client bottleneck before we reach 1 million txns and leaves this seems reasonable, in the future we can do a rust implementation of fd_bmtree to further optimize this if necessary.

[aeyakovenko]

should this be closed since #64 is merged? @anoushk1234

[anoushk1234]

should this be closed since #64 is merged? @anoushk1234

@aeyakovenko 64 only defines the receipt data structure, 52 adds it to the bankhash.

[0xSol]

Per current SIMD process, If a proposal reaches 6 months without activity, the proposal will be marked as stale to be closed. A new proposal can be opened if the proposal is closed and has a chance of reaching consensus. Will wait until next week to close if no additional updates.