Tage way tuning align

[CJ362ff]

Summary by CodeRabbit

• Refactor

  • Updated branch predictor defaults to increase capacity and vary associativity across internal tables for improved coverage.

• Performance

  • Expect improved prediction accuracy and reduced misprediction penalties from the retuned predictor defaults.

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

📝 Walkthrough

Walkthrough

BTBTAGE default per-table configuration in src/cpu/pred/BranchPredictor.py changed: tableSizes updated to [2048, 2048, 8192, 2048, 2048, 8192, 8192, 2048] and numWays updated to [2, 2, 4, 4, 4, 2, 2, 2]. No other logic or control flow modified.

Changes

Cohort / File(s)	Summary
Branch Predictor Configuration src/cpu/pred/BranchPredictor.py	Updated BTBTAGE.tableSizes default from eight 2048 entries to [2048, 2048, 8192, 2048, 2048, 8192, 8192, 2048]. Updated BTBTAGE.numWays default from [2]*8 to [2, 2, 4, 4, 4, 2, 2, 2]. No other parameters or code paths changed.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• cpu-o3: update tage index #627: Modifies BTBTAGE configuration in the same file (src/cpu/pred/BranchPredictor.py).
• cpu-o3: Refactor numWays to support per-table configuration #793: Refactors per-table numWays/tableSizes handling for BTBTAGE.
• Restore/per table numways #798: Adjusts/restores per-table numWays and tableSizes behavior related to these defaults.

Suggested labels

align-kmhv3, perf

Suggested reviewers

• jensen-yan
• Yakkhini

Poem

🐰 I hopped through rows of bits tonight,
Tuned tables wide to get predictions right.
Ways and sizes set with care,
I hide my guesses in each lair.
Hop, nibble, cache—silent delight.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Title check	❓ Inconclusive	The title is vague and lacks clarity about the specific changes; 'align' is non-descriptive and does not convey meaningful information about what is being tuned or aligned.	Revise the title to be more specific and descriptive, such as 'Tune BTBTAGE table sizes and associativity for improved performance' or 'Increase BTBTAGE table sizes and ways for tables T2, T4-T6'.

✅ 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 unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch tage-way-tuning-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)	2.2665	-
This PR	2.2655	📉 -0.0010 (-0.05%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: f9e128f
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.28	18.25	+0.13 🟢

[Generated by GEM5 Performance Robot]
commit: f9e128f
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.28	18.30	-0.10 🔴

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2579	📉 -0.0086 (-0.38%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: 8bdbac9
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.27	18.25	+0.09 🟢

[Generated by GEM5 Performance Robot]
commit: 8bdbac9
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.27	18.28	-0.05 🔴

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

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

Inline comments:
In `@src/cpu/pred/BranchPredictor.py`:
- Around line 1047-1055: The change to the base params (tableSizes and numWays
in BranchPredictor.py) unintentionally increases BTBTAGE capacity; either
restore conservative defaults on the base class or isolate the larger geometry
into a dedicated subclass: if your intent is an 8-way ×8K-entry BTBTAGE, set
tableSizes = VectorParam.Unsigned([1024] * 8, ...) while keeping numWays =
VectorParam.Unsigned([8] * 8, ...); otherwise revert tableSizes and numWays in
the base BranchPredictor default to the previous smaller values (or create a new
BTBTAGE-specific subclass used by BTBTAGE(), DecoupledBPUWithBTB and
BTBTAGEUpperBound that overrides tableSizes/numWays) so the global default does
not inflate all BTB instances.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: cbe60c45-94ab-46c6-8211-546c8550797c

📥 Commits

Reviewing files that changed from the base of the PR and between 8bdbac9 and c113f1c.

📒 Files selected for processing (1)

• src/cpu/pred/BranchPredictor.py

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2975	📈 +0.0310 (+1.37%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: c113f1c
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.56	18.25	+1.68 🟢

[Generated by GEM5 Performance Robot]
commit: c113f1c
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.56	18.27	+1.59 🟢

[coderabbitai]

🧹 Nitpick comments (1)

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

1047-1055: LGTM — heterogeneous BTBTAGE configuration looks intentional and performance-validated.

The new configuration increases total capacity from ~32K to ~123K entries (~3.75x) with per-table tuning:

• Tables T2, T5, T6 scaled to 8192 sets (13-bit index) vs. 2048 (11-bit) for others
• Associativity varies: T3 at 8 ways, T0/T2/T4/T6 at 4 ways, T1/T5/T7 at 2 ways

The C++ constructor already handles heterogeneous tableSizes and numWays correctly (calculates tableIndexBits[i] per-table via ceilLog2(tableSizes[i])).

Optional: Consider adding a brief inline comment documenting the tuning rationale (e.g., RTL alignment, memory budget target) for future maintainability.

,

🤖 Prompt for AI Agents

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

In `@src/cpu/pred/BranchPredictor.py` around lines 1047 - 1055, Add a short inline
comment above the heterogeneous BTB/TAGE configuration (near tableSizes,
TTagBitSizes, TTagPcShifts, histLengths, numWays) explaining the tuning
rationale (e.g., which tables were enlarged to 8192 sets, associativity choices
like T3=8 ways, and the target memory/capacity/RTL alignment trade-offs) so
future readers understand why these specific sizes/ways were chosen; keep the
comment concise (1–2 sentences) and place it adjacent to the existing parameter
block for maintainability.

🤖 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`:
- Around line 1047-1055: Add a short inline comment above the heterogeneous
BTB/TAGE configuration (near tableSizes, TTagBitSizes, TTagPcShifts,
histLengths, numWays) explaining the tuning rationale (e.g., which tables were
enlarged to 8192 sets, associativity choices like T3=8 ways, and the target
memory/capacity/RTL alignment trade-offs) so future readers understand why these
specific sizes/ways were chosen; keep the comment concise (1–2 sentences) and
place it adjacent to the existing parameter block for maintainability.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 622f1e29-5ea9-49e9-a9ff-b8c6ccd4b251

📥 Commits

Reviewing files that changed from the base of the PR and between c113f1c and ac20e3f.

📒 Files selected for processing (1)

• src/cpu/pred/BranchPredictor.py

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2700	📈 +0.0035 (+0.15%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: ac20e3f
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.47	18.25	+1.16 🟢

[Generated by GEM5 Performance Robot]
commit: ac20e3f
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.47	18.56	-0.51 🔴

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2653	📉 -0.0012 (-0.06%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: 0c52707
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.45	18.25	+1.09 🟢

[Generated by GEM5 Performance Robot]
commit: 0c52707
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.45	18.47	-0.08 🔴

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2987	📈 +0.0322 (+1.42%)

✅ Difftest smoke test passed!

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.3043	📈 +0.0378 (+1.67%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: 2d970fc
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.51	18.25	+1.42 🟢

[Generated by GEM5 Performance Robot]
commit: 2d970fc
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.51	18.44	+0.38 🟢

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2676	📈 +0.0011 (+0.05%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: fe5af39
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.21	18.25	-0.23 🔴

[Generated by GEM5 Performance Robot]
commit: fe5af39
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.21	18.51	-1.63 🔴

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2839	📈 +0.0174 (+0.77%)

✅ Difftest smoke test passed!

[github-actions]

🚀 Coremark Smoke Test Results

Branch	IPC	Change
Base (xs-dev)	2.2665	-
This PR	2.2847	📈 +0.0182 (+0.80%)

✅ Difftest smoke test passed!

[XiangShanRobot]

[Generated by GEM5 Performance Robot]
commit: 0f712f2
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Master	Diff(%)
Score	18.33	18.25	+0.43 🟢

[Generated by GEM5 Performance Robot]
commit: 0f712f2
workflow: gem5 Align BTB Performance Test(0.3c)

Align BTB Performance

Overall Score

	PR	Previous Commit	Diff(%)
Score	18.33	18.21	+0.67 🟢