We have some language about how timestamps should be coarsened in our spec:

WebGPU also specifies the "timestamp-query" feature, which provides high precision timing of GPU operations. The feature is optional, and a WebGPU implementation may limit its exposure only to those scenarios that are trusted. Alternatively, the timing query results could be processed by a compute shader and aligned to a lower precision.
https://gpuweb.github.io/gpuweb/#security-timing

I think we should probably point to the High Resolution Time spec to be more prescriptive here.
https://w3c.github.io/hr-time/#dfn-coarsen-time