Device wall
Thirty-seven families, biased toward the SKUs Thai retail actually sells. Flagships are present; they are not the default. Each phone has a soak log: last charge, last thermal run, last OS image.
Benchmark lab
The wall, the soak, and the awkward handoff
Mobile Performance Benchmark Analytics only holds if the room that produces the traces looks a little like the city outside. This page is the lab, not the brochure.
Network fixtures
Captures from commute loops in Wattana, a river-side 5G-to-4G hop, and hotel Wi‑Fi that advertises high bandwidth. Fixtures are versioned. We retire them when a carrier changes behaviour.
Thermal practice
A 12-minute bag soak before jank runs. Air-conditioning is a variable we record, not a luxury we pretend away. Module 03 of Field Benchmark Mastery is taught against this ritual.
What we instrument, and what we refuse
We capture CPU, GPU, radio, and frame timing. We do not publish synthetic scores from a single cool device as “the” number. We do not pretend Unity or Unreal game loops are in scope — see the flagship FAQ.
Every teaching trace in the archive carries a header: device family, OS build, fixture version, ambient notes. If a header is missing, the trace is not used in class.
See how courses use the labVisit during Studio Residency
Open lab hours are not a walk-in gallery. Device wall time is booked with a residency seat or by writing to the desk. If you only need a fixture recipe, Trace Reader notes may be enough.
Ask about lab time