Physical Apple Silicon release acceptance · Noisedeck → Syphon
This report covers separate WebGL2 and WebGPU tests at 1024 × 1024. Each test requests 20 seconds of data from Noisedeck to Syphon. These tests do not establish 1920×1080 camera performance or sustained delivery at 60 distinct frames per second. They do not measure physical display output or stability over several hours. Render FPS measures the browser render rate. The receiver counts unique latency markers. This count does not measure delivered FPS. Sender counters contain cumulative totals. Their intervals differ from the intervals that determine the reported rates. The backend tests ran separately. They do not provide a simultaneous comparison.
| Backend | Output | Duration | Render FPS | Unique latency markers | Reported GPU |
|---|---|---|---|---|---|
| WebGL2 | 1024 × 1024 | 20 s requested / 20.276 s observed | 60.021 | 588 | Apple M4 |
| WebGPU | 1024 × 1024 | 20 s requested / 20.050 s observed | 59.994 | 600 | Not reported |
The test logs contain no timeline. This report does not reconstruct one.
The test does not report the exact computer model. The report identifies the GPU only when the renderer name contains a recognized model. Otherwise, it reports no GPU model. The acceptance check requires browser measurements for at least 90% of the requested duration.
Source workflow run · Measurement JSON · Artifact checksums
b94d82e9bd56b84d8fe7e576ab09643e0b9cab2f95b737025c03e5a8d4e42e3aec1e8cc35c3d1efa539ee0892f575361195da69ca46abc2996488cd3e1feefa0a1188e70aae0169fdc1e7901cc8cc939Noisemaker version: 1.0.118. The report.json file contains hashes of the source inputs. It excludes private logs and other environment text.