[consensus] parallelize verify opt proposal

[ibalajiarun]

Description

These fixes get about 1ms block time improvement in Devnet.

• Parallelize expensive computation in OptProposalMsg::verify.
• A minor fix to batch fetch logic to subscribe to payload only if necessary.

[github-actions]

✅ Forge suite realistic_env_max_load success on 0844c55d827d46f3822ab3e74b0ca80d8dd470c9

two traffics test: inner traffic : committed: 12579.44 txn/s, latency: 3000.30 ms, (p50: 3000 ms, p70: 3000, p90: 3300 ms, p99: 4000 ms), latency samples: 4783040
two traffics test : committed: 99.96 txn/s, latency: 2389.49 ms, (p50: 2300 ms, p70: 2800, p90: 3100 ms, p99: 3900 ms), latency samples: 1740
Latency breakdown for phase 0: ["MempoolToBlockCreation: max: 1.785, avg: 0.784", "ConsensusProposalToOrdered: max: 0.165, avg: 0.163", "ConsensusOrderedToCommit: max: 0.673, avg: 0.501", "ConsensusProposalToCommit: max: 0.837, avg: 0.665"]
Max non-epoch-change gap was: 1 rounds at version 49948 (avg 0.00) [limit 4], 1.26s no progress at version 2468069 (avg 0.07s) [limit 15].
Max epoch-change gap was: 0 rounds at version 0 (avg 0.00) [limit 4], 0.98s no progress at version 2251698 (avg 0.98s) [limit 16].
Test Ok

• Grafana dashboard
• Humio Logs
• Axiom Logs
• Validator CPU Profile
• Fullnode CPU Profile
• Test runner output
• Test run is land-blocking