Utage improve align

[CJ362ff]

Summary by CodeRabbit

• Chores
  • Updated branch predictor configuration to improve prediction accuracy and overall simulation performance.
  • Removed a deprecated configuration option and cleaned up minor formatting for clarity.

[coderabbitai]

📝 Walkthrough

Walkthrough

Updated MicroTAGE public parameters in the branch predictor: tag bit sizes reduced, history lengths changed, associativity increased, and a base table size parameter removed; small formatting adjustments only.

Changes

Cohort / File(s)	Summary
MicroTAGE Parameter Updates src/cpu/pred/BranchPredictor.py	Changed TTagBitSizes from [16, 16, 16, 16] to [13, 13, 13, 13]; updated histLengths from [5, 9, 17, 27] to [5, 17, 56, 109]; changed numWays from 1 to 2; removed the public baseTableSize parameter; minor formatting/spacing edits.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Suggested reviewers

• Yakkhini
• tastynoob

Poem

🐰 I nibble bits and trim the tags,

Histories lengthen, two ways to wag,
A tiny param hopped away,
Predicting hops from night to day,
🥕✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Title check	❓ Inconclusive	The title 'Utage improve align' is vague and does not clearly describe the actual changes made to the codebase.	Revise the title to be more descriptive of the specific changes, such as 'Update MicroTAGE branch predictor parameters' to accurately reflect the modifications to TTagBitSizes, histLengths, numWays, and baseTableSize.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ 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 utage-improve-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.}

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2607	-
This PR	2.2625	📈 +0.0018 (+0.08%)

✅ Difftest smoke test passed!

[coderabbitai]

🧹 Nitpick comments (1)

src/cpu/pred/BranchPredictor.py (1)

1082-1082: Please cover the new 2-way MicroTAGE path.

Line 1082 now enables way == 1 allocation/update behavior in src/cpu/pred/btb/microtage.cc:518-527. A focused regression or config that forces second-way hits/replacement would make this tuning safer to land.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/cpu/pred/BranchPredictor.py` at line 1082, Branch predictor changes
enable the MicroTAGE two-way path (allocation/update for way == 1) but there's
no focused test exercising it; add a small regression/config that forces
second-way hits and replacements to validate the new path. Create a
unit/regression harness that instantiates the BranchPredictor/MicroTAGE with
numWays=2 (the Param.Unsigned setting in BranchPredictor.py) and run a synthetic
branch stream that causes collisions in the BTB to trigger allocations/updates
on way==1 (the code path in microtage.cc handling way == 1); ensure the test
asserts expected predictions/updates and fails if way==1 never executes so the
two-way behavior is covered.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@src/cpu/pred/BranchPredictor.py`:
- Line 1082: Branch predictor changes enable the MicroTAGE two-way path
(allocation/update for way == 1) but there's no focused test exercising it; add
a small regression/config that forces second-way hits and replacements to
validate the new path. Create a unit/regression harness that instantiates the
BranchPredictor/MicroTAGE with numWays=2 (the Param.Unsigned setting in
BranchPredictor.py) and run a synthetic branch stream that causes collisions in
the BTB to trigger allocations/updates on way==1 (the code path in microtage.cc
handling way == 1); ensure the test asserts expected predictions/updates and
fails if way==1 never executes so the two-way behavior is covered.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: d7df1b22-069c-4f7b-982b-64d85e045d02

📥 Commits

Reviewing files that changed from the base of the PR and between 84ec451 and 1198737.

📒 Files selected for processing (1)

• src/cpu/pred/BranchPredictor.py

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2607	-
This PR	2.2574	📉 -0.0034 (-0.15%)

✅ Difftest smoke test passed!

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2607	-
This PR	2.2692	📈 +0.0084 (+0.37%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: 1198737
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.26	18.27	-0.02 🔴