feat(ci): add memory and GC metrics to benchmark comparison (#69)

## Summary
- Enhance benchmark comparison to show memory allocation and GC
collection metrics
- Add comprehensive metric comparison in PR comments

## Changes
1. **Median Baseline Calculation**
- Calculate median values for execution time, memory, and GC collections
from last 10 runs
   - More stable baseline that reduces noise from outliers

2. **Enhanced Comparison Table**
   - Add Memory (MB) column with baseline and percentage diff
   - Add Gen0, Gen1, Gen2 columns showing GC collection counts
   - Color-coded indicators for both time and memory changes

3. **Improved Documentation**
   - Update footer to explain all metrics
   - Clarify baseline methodology

## Test Plan
- [x] Workflow syntax validated
- [ ] Wait for benchmark workflow to run and verify output format
- [ ] Confirm memory and GC metrics appear in PR comment

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Sonnet 4.5 <[email protected]>

Author: eric-wang-1990