[mono-vm][specializer] Move a fusing operation from lowering to destacking phase

[vineethk]

Description

There was a mismatch where cmp+branch fusion was performed, because those micro-ops came later.

In this PR, we move this fusion from lowering to the destacking phase.

It results in fewer stackless exec IR instructions in some cases, and fewer slots (e.g., on the fibonacci example).

How Has This Been Tested?

Existing tests have been upgraded.

Type of Change

• Refactoring

---

Note

Medium Risk
Touches control-flow semantics across IR fusion, lowering, and runtime execution/verification; errors here could miscompile branches or change loop termination behavior despite test updates.

Overview
Moves compare+branch fusion out of lowering and into the destacking/SSA fusion pipeline by introducing CmpOp, representing comparisons as BinaryOp::Cmp(_), and fusing BinaryOp{Cmp} + BrTrue/BrFalse into new IR terminators BrCmp / BrCmpImm.

Updates lowering to consume BrCmp* directly (and to lower standalone BrFalse), adding new micro-ops JumpGreaterU64Imm and JumpLessEqualU64Imm plus corresponding interpreter, verifier, CFG/gas, and display support. Golden tests are refreshed to reflect fewer temporary slots and newly-lowered control-flow cases that were previously skipped.

^{Reviewed by Cursor Bugbot for commit 2f56ed6. Bugbot is set up for automated code reviews on this repo. Configure here.}

[vineethk]

• [mono-move] Migrate from using the old type representation to the new one #19530
• [mono-vm][specializer] Remove explicit types from simulated operand stack #19466
• [mono-vm] Factor specializer pipeline out of the executable and into the specialize crate. #19454
• [mono-vm][specializer] Move a fusing operation from lowering to destacking phase #19431 👈 (View in Graphite)
• [mono-vm] Representation and perf refactoring within specializer #19427
• Shared data structures and algorithms crate #19388
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 2712de9. Configure here.}

[github-actions]

✅ Forge suite compat success on ca049383dd80675149ef2d0042668964f9f9107a ==> 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749

Compatibility test results for ca049383dd80675149ef2d0042668964f9f9107a ==> 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749 (PR)
1. Check liveness of validators at old version: ca049383dd80675149ef2d0042668964f9f9107a
compatibility::simple-validator-upgrade::liveness-check : committed: 14316.68 txn/s, latency: 2406.55 ms, (p50: 2400 ms, p70: 2700, p90: 3100 ms, p99: 3700 ms), latency samples: 470880
2. Upgrading first Validator to new version: 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749
compatibility::simple-validator-upgrade::single-validator-upgrade : committed: 6276.21 txn/s, latency: 5384.34 ms, (p50: 5900 ms, p70: 6000, p90: 6100 ms, p99: 6200 ms), latency samples: 217100
3. Upgrading rest of first batch to new version: 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749
compatibility::simple-validator-upgrade::half-validator-upgrade : committed: 6408.09 txn/s, latency: 5299.79 ms, (p50: 5800 ms, p70: 5900, p90: 6000 ms, p99: 6200 ms), latency samples: 220500
4. upgrading second batch to new version: 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749
compatibility::simple-validator-upgrade::rest-validator-upgrade : committed: 10940.31 txn/s, latency: 2960.77 ms, (p50: 3100 ms, p70: 3300, p90: 3500 ms, p99: 3700 ms), latency samples: 359120
5. check swarm health
Compatibility test for ca049383dd80675149ef2d0042668964f9f9107a ==> 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749 passed
Test Ok

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

[github-actions]

✅ Forge suite realistic_env_max_load success on 2f56ed63d42c7cbc8cb6f5db12c1a5a9d6713749

two traffics test: inner traffic : committed: 15712.80 txn/s, latency: 1089.75 ms, (p50: 1000 ms, p70: 1100, p90: 1200 ms, p99: 1500 ms), latency samples: 5869280
two traffics test : committed: 100.01 txn/s, latency: 822.89 ms, (p50: 800 ms, p70: 900, p90: 1000 ms, p99: 1200 ms), latency samples: 1760
Latency breakdown for phase 0: ["MempoolToBlockCreation: max: 0.292, avg: 0.248", "ConsensusProposalToOrdered: max: 0.114, avg: 0.111", "ConsensusOrderedToCommit: max: 0.194, avg: 0.175", "ConsensusProposalToCommit: max: 0.306, avg: 0.286"]
Max non-epoch-change gap was: 2 rounds at version 48782 (avg 0.00) [limit 4], 2.15s no progress at version 48782 (avg 0.06s) [limit 15].
Max epoch-change gap was: 0 rounds at version 0 (avg 0.00) [limit 4], 0.21s no progress at version 2731760 (avg 0.21s) [limit 16].
Test Ok

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