Cardano Updates

The sequential engine was never quantizing timestamps, so each event had a
unique timestamp and rayon parallelism never kicked in during TX-only periods.

Two independent optimizations, each configurable separately:
- Timestamp quantization: now applied to all events in the sequential engine,
  controlled by existing timestamp-resolution-ms (turbo.yaml sets 1.0ms)
- TX batch generation: new tx-batch-window-ms config option batches all TX
  generation events within a time window into one timestamp (turbo.yaml sets
  10ms). Independent of timestamp resolution.

Also lowers parallel-threshold from 10 to 4 in turbo.yaml.

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

UI: Add graph layout modes (force, circular, Mercator)

Track the maximum tx backlog length (overflow queue behind the mempool)
via events, and add a leios-tx-backlog-max-size config parameter that
applies backpressure on the tx generator when the backlog is full.
Peer transactions are unaffected by the cap.

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

Extract the hardcoded PARALLEL_THRESHOLD constant into a configurable
parallel-threshold parameter (default 10, was 32). Add engine and
parallel-threshold to config.schema.json. Disable rayon in the
determinism test since event channel ordering is non-deterministic
under parallel execution.

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

The sequential engine does not run Actor-based attackers
(EBWithholdingAttacker). Log a warning so the config isn't
silently ignored.

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

- Extract actor-model simulation (NodeListWrapper, ActorSimulation,
  init_nodes, run logic) into sim/actor.rs
- sim.rs is now a thin dispatch layer: Simulation newtype wrapping
  SimulationInner enum (Actor vs Sequential)
- Unify sequential single-shard and multi-shard builders into a
  single build_typed() function
- Group cross-shard state into CrossShardState sub-struct
- build() takes event_sender directly, creates its own infrastructure
- Remove dead code (per_shard_node_configs, init_node_impls)
- Net -407 lines across sim.rs + sequential.rs

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

- Add ShardStrategy enum (round-robin, zero-latency-clusters) configurable
  via shard-strategy in YAML parameters
- Extract sharding logic into sim-core/src/sharding/ module with pluggable
  strategy files
- zero-latency-clusters uses union-find to keep 0-latency-connected nodes
  on the same shard, improving CMB lookahead
- Log shard sizes after assignment for balance diagnostics
- Restore single-shard select! fast path (avoids tokio::spawn overhead
  that caused 50% regression vs pre-sharding baseline)
- Update config schema and test parameter files

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>