[Alex] PolyBench Hardware Baseline Collection for H5 Validation

[syifan]

Priority: CRITICAL (H5 Milestone)

Issue: Missing PolyBench hardware baselines are blocking H5 milestone validation

Context:

• H5 milestone claims 15+ intermediate benchmarks with 13.3% accuracy
• Reality: Only 7 microbenchmarks have accuracy data
• MISSING: PolyBench intermediate benchmarks lack M2 hardware baselines entirely

Required Work:

1. Collect M2 hardware baseline data for 7 PolyBench benchmarks
2. Use cache-disabled configuration for accurate timing baselines
3. Apply linear regression methodology for startup overhead separation
4. Validate baseline data quality (statistical significance)
5. Update accuracy calibration framework with new baselines

Success Criteria:

• All 7 PolyBench benchmarks have verified M2 hardware baselines
• Baselines pass statistical validation (R² >95% where applicable)
• Framework ready for accuracy calibration execution

Dependencies:

• Issue [Alex] -> [Alex] URGENT: Accuracy Data Validation - CI-Generated Results Analysis #506 accuracy validation results
• CI infrastructure stability (recently fixed)

Timeline: URGENT - H5 validation depends on this data