cpu-o3: Add replayed load completion counter

[happy-lx]

Add a load instruction replay success rate counter to evaluate the effectiveness of replay behavior and provide a reference for RTL.

Change-Id: I78887a7eb4d90ad5b5a3a870977ff9aabd621468

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

Configuration used: defaults

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 7542d58 and 014d5f3.

📒 Files selected for processing (3)

• src/cpu/o3/dyn_inst.hh
• src/cpu/o3/lsq_unit.cc
• src/cpu/o3/lsq_unit.hh

---

📝 Walkthrough

Walkthrough

This PR introduces instruction replay tracking by adding an everReplayed() accessor to DynInst and implementing three new LSQUnit statistics to monitor replayed load execution: issue-to-pipeline count, ready-to-finish count, and their computed finish rate. The statistics are initialized and incremented at relevant stages of the load pipeline.

Changes

Cohort / File(s)	Summary
Instruction Replay Metadata src/cpu/o3/dyn_inst.hh	Added boolean flag and public everReplayed() const accessor to track whether an instruction has ever been replayed.
LSQUnit Statistics Definition src/cpu/o3/lsq_unit.hh	Added three public statistics members to LSQUnitStats: replayedLoadIssueToPipe (Scalar), replayedLoadReadyToFinish (Scalar), and replayedLoadFinishRate (Formula/Ratio).
LSQUnit Statistics Implementation src/cpu/o3/lsq_unit.cc	Initialized new statistics with precision setting; added counter increments in issueToLoadPipe() and executeLoadPipeSx() to track replayed load progression through the pipeline.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Possibly related PRs

• cpu-o3: Add replay-based memdep prediction option #720: Modifies DynInst class with replay-related fields and accessors in the same header file.
• cpu-o3: fix rar raw replay #611: Adjusts load replay logic in LSQUnit's executeLoadPipeSx and related replay handling paths.

Suggested reviewers

• Ergou-ren
• jensen-yan

Poem

🐰 A load replays, we track its way,
From pipe to finish, day by day,
With ratios and counters bright,
The replay story now in sight! 📊✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main change: introducing replayed load completion counters/statistics across three files in the O3 CPU module.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch lsq-replay-counter

Tip

Try Coding Plans. Let us write the prompt for your AI agent so you can ship faster (with fewer bugs).
Share your feedback on Discord.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.1727	-
This PR	2.1727	➡️ 0.0000 (0.00%)

✅ Difftest smoke test passed!

[happy-lx]

Detailed stats:
replay-successful-rate.xlsx

According to SPEC06：