configs: Disable MGSC for SC benefit ablation

[jensen-yan]

Summary

• Disable the whole MGSC predictor in the KMHV3 DecoupledBPUWithBTB config.
• Keep individual SC table settings untouched; this PR only measures whole-SC contribution on current xs-dev.

Motivation

PR #852 previously showed about a 0.3 Int-score drop when SC was effectively removed. Since current mainline includes the recent PHR recovery fix, rerun the simplest SC-off ablation first before deciding whether table-level ablations are worth repeating.

Validation

• python3 -m py_compile configs/example/kmhv3.py
• git diff --check

Expected follow-up: compare the perf robot score against current xs-dev.

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 8a79100b-5e30-4659-bf3a-236279f011d0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/sc-disable-ablation-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)	2.2906	-
This PR	2.1875	📉 -0.1031 (-4.50%)

✅ Difftest smoke test passed!

[jensen-yan]

Closing this draft because the head branch is ; the workflow only matches slash-free push branches. Replacement PR: #883 (), which has triggered the perf workflow.

[jensen-yan]

Correction: the old head branch was codex/sc-disable-ablation-align, which does not match the gem5 Align BTB Performance Test branch pattern *-align because the branch name contains a slash. Replacement PR is #883 on sc-disable-ablation-align, and it has triggered the 0.3c perf workflow: https://github.com/OpenXiangShan/GEM5/actions/runs/26746954461