Microtage abtb align

[jensen-yan]

test close microTage and test close ABTB

Summary by CodeRabbit

• Chores
  • Disabled branch predictor subcomponents in a specific configuration variant.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 71848b82-490a-4db0-a152-c5f7d288b565

📥 Commits

Reviewing files that changed from the base of the PR and between 11bcbe6 and 17f6da9.

📒 Files selected for processing (1)

• configs/example/kmhv3.py

---

📝 Walkthrough

Walkthrough

The pull request disables two branch predictor subcomponents (ABTB and microtage) within the DecoupledBPUWithBTB configuration path in the KMH V3 system parameters. Their enabled flags are toggled from True to False in the branch predictor configuration.

Changes

Cohort / File(s)	Summary
Branch Predictor Configuration configs/example/kmhv3.py	Disabled ABTB and microtage subcomponents by setting their enabled flags to False within the DecoupledBPUWithBTB configuration block.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly related PRs

• Add kmhv3 #609: Modifies branch predictor configuration in the same DecoupledBPUWithBTB path, adding or simplifying BP handling logic.
• Test align #612: Touches the same ABTB and microtage flags in the KMH V3 branch predictor configuration.
• Resolve align #621: Modifies DecoupledBPUWithBTB branch predictor setup, including changes to the abtb.enabled flag.

Suggested reviewers

• tastynoob
• Yakkhini

Poem

🐰 A gentle tweak to configs so neat,
Branch paths refined, the logic complete,
ABTB and microtage rest at ease,
While DecoupledBPU hops with grace and wheeze! ✨

🚥 Pre-merge checks | ✅ 3 | ❌ 2

❌ Failed checks (1 warning, 1 inconclusive)

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.
Title check	❓ Inconclusive	The title 'Microtage abtb align' is vague and does not clearly describe the actual changes, which are disabling these components.	Consider using a more descriptive title such as 'Disable microTAGE and aBTB in kmhv3 configuration' to clearly convey the main change.

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

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

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch microtage-abtb-align

---

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.}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 17f6da93e4

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Keep ABTB and MicroTAGE enabled by default

This change hard-disables both abtb and microtage for every DecoupledBPUWithBTB run in kmhv3.py, which silently changes the default Kunminghu v3 frontend behavior and can significantly skew branch-prediction-sensitive performance experiments. Because there is no new CLI/config switch to opt back in, users running the standard config cannot reproduce prior baselines or compare against the intended composed predictor stack.

Useful? React with 👍 / 👎.

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.3051	-
This PR	2.2602	📉 -0.0449 (-1.95%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: 17f6da9
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	17.74	18.70	-5.14 🔴

[Generated by GEM5 Performance Robot]
commit: 17f6da9
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	17.74	17.75	-0.02 🔴