Skip to content

Submission cadence and delivered reports

The campaign requests 125, 1,000 and 10,000 reports per second using a high-resolution waitable timer where supported. It records actual QPC submission intervals, SDK return codes and independently observed content. Per-run results include actual rate, missing reports and submission cost.

A requested 10 kHz submission schedule is not proof of 10 kHz physical USB polling. The virtual full-speed HID descriptor requests a 1 ms interrupt interval. Windows scheduling, buffers and virtual USB behavior remain part of the path. These finite-rate scenarios do not establish an absolute maximum sustainable throughput.

See the measured table for all rates and the raw CSV archive to reconstruct the observed cadence. Longer saturation and multi-device campaigns must be executed before claiming a maximum throughput or long-term stability bound.