Revision · Project 7 — Solana Mini-Chain

Project 7 built solmini, a Solana-style mini-runtime aimed at raw throughput: a verifiable clock and a scheduler that runs non-conflicting transactions in parallel across threads. It closes the playbook by mapping all three chains onto real Anchor. Here’s what to carry forward.

What this part covered

• The throughput problem — Ethereum serializes transactions because a contract’s storage footprint is unknown until you run it; Solana’s narrow bet is to optimize almost everything for transactions per second per dollar, accepting the trilemma trade-off (hardware, centralization pressure) to get there.
• Proof of History — a sequential SHA-256 hash chain is a clock you can’t fake without doing the work; producing it is serial, but anyone can verify it in parallel by recomputing, and flipping one byte makes verification fail. A clock replaces a conversation between validators about when things happened.
• The account model — state lives outside code in typed accounts, and programs are stateless pure functions over the accounts handed to them; because code holds no hidden storage, a transaction’s footprint is declarable up front.
• Declared footprints — each transaction lists the accounts it reads and writes, so the scheduler can see which transactions can’t interfere without running anything — “shared XOR mutable” lifted from variables to accounts.
• The Sealevel scheduler — dependency layering groups transactions into conflict-free batches, and each batch runs across threads with rayon’s par_iter_mut, which hands each worker a disjoint &mut, so the runtime can’t race by construction.
• Honest trade-offs — the scheduler’s O(n^2) conflict scan and SHA-256 keys are deliberate simplifications, and the demo’s speedup varies with core count and isn’t a benchmark.
• The Anchor bridge — the capstone maps solmini onto real Solana/Anchor (discriminators, Borsh, rent, CPI, program-derived addresses), so the toy runtime becomes your mental model for production programs.

The takeaway

The recurring question sharpens to fearless parallelism at scale: rayon’s par_iter_mut and your scheduler enforce the same “shared XOR mutable” guarantee at two levels, so the data race that “just run these on threads” invites in C is again unrepresentable. Day 30 sets all three mini-chains side by side and answers, in code you wrote, why Bitcoin, Ethereum, and Solana each exist — then points you at real Anchor development.