Tide Benchmarks

Real performance numbers from ACS Dashboard (11 pages, 12 Hermes gateways, Go binary).

⚠️ Timing numbers (revisit/wsDom/warmPaint/coldPaint) are PROVISIONAL — pending vanilla-baseline re-validation by the benchmark agent. Bundle size is final (Tier-1).

Client-Side Performance

Revisit render (sessionStorage) 0ms
WS → DOM update 12ms
First paint (server cache warm) 50ms
First paint (cold, no cache) 300ms
Bundle size (gzipped, core+skeleton) 2.8KB

API Performance (Server-Side Cache)

Server-side cache pattern (Go/Fiber). Before = cold compute, After = warm cache.

Endpoint Before (ms) After (ms) Improvement
/api/dashboard 8,941 6 1,490x
/api/gateways 1,828 2 914x
/api/stack/status 5,000 6 833x
/api/agentic/tooling-status 5,609 16 350x
/api/agentic/summary 200 10 20x
/api/announcements 50 8 6x
/api/daemon/status 10 4 2.5x

1,490× — fastest improvement (the /api/dashboard endpoint). This reflects the Go/Fiber server-side cache pattern, not pure Tide client speed.

Methodology

Benchmarks were run against the ACS Dashboard production app — a SolidJS application with 11 pages and 12 Hermes gateways, served by a Go/Fiber binary.

Automated testing via CloakBrowser. Client-side timing measured via Performance API. Server-side timing measured from Go middleware logging.

Timing numbers are provisional (pending Tier-2 re-validation). Bundle size of 2.8KB gzipped is final.