Add Yoga benchmark comparisons

[hyfdev]

Summary

• extend the benchmark runner from one hard-coded target pair to independent Taffy API and Yoga API comparison groups
• add modeled chat, 1,000-leaf wide-tree, and 511-node deep-tree transactions for @taffyjs/yoga, yoga-layout@3.2.1, and @taffyjs/yoga-wasm
• render every comparison group on the benchmark page with a group-local baseline and no cross-API ranking

Why

The Node and Wasm benchmark established the shared end-to-end measurement pipeline. This change reuses that pipeline for the shipped Yoga-compatible packages so native Taffy Yoga, upstream Yoga, and Taffy Yoga Wasm can be compared through the same public Yoga API transactions.

Validation

• ran all three Yoga scenarios across all three targets with equal output checksums
• ran vp run check
• ran vp run check:wasm with an isolated temporary directory
• built the website against both the retained Taffy-only dataset and a Yoga-only local dataset
• completed a bounded adversarial review with no blocking findings

Published results

This PR intentionally does not replace benchmarks/results/published.json. The retained results were produced on an Apple M3 Pro, while this implementation was validated on Linux. Run vp run benchmark:update-website on the benchmark host before publishing the Yoga numbers.

[Copilot]

The job was not started because recent GitHub Actions payments have failed or your spending limit needs to be increased.