sparkDash 2.0 open-sourced: a factory-style dashboard for running private GPU inference clusters
aigclink · x · 2026-10-10
MiaAI Lab open-sourced sparkDash 2.0 (Apache-2.0/MIT, 601 GitHub stars), a multi-unit monitoring dashboard that treats a private inference cluster like a factory floor. It answers four core questions: what each GPU is doing (tokens/sec, decode vs prefill, power and cost per 1K tokens), per-machine status cards covering GPU/CPU/memory/storage/network and the served model (any NVIDIA Linux box works, one card per GPU), quality drift after switching models/quantization/engines, and per-token unit cost. Built from the team's real needs running GLM 5.3 Flash on 4 DGX Spark units—a solid reference for teams pursuing on-prem compute and model sovereignty.
More from Infra
- Building a pit crew for Grok Bot: frontier model plans, free models grind — alexcovo_eth · 2026-10-10
- AI boom turns into a debt boom: Oracle 5y CDS near record 261bps, implying 20.4% default odds — cyb3rops · 2026-10-10
- Fireworks AI Discloses Security Incident Involving Unauthorized Use of Internal Credentials — lqiao · 2026-10-10
- US grid adds 86GW this year while AI labs need hundreds of GW of power — FinanceYF5 · 2026-10-10
- Running Qwen3.6 35B-A3B with 131K context and vision on a 6GB RTX 2060 — full config — Szadbaverem69 · 2026-10-10
- SpaceX building industrial base to make up to 1,000 Starships a year, with Florida GigaBay 11x bigger than current Megabay — XFreeze · 2026-10-10