Abstract
Techniques are described for instant performance classification of a computing device using an aggregate CPU frequency score. A client device obtains, for each CPU core, a maximum clock frequency value from an operating system-exposed hardware descriptor and computes a continuous score by aggregating the per-core maximum frequencies, such as by summation. The score can be computed with low overhead and reported in telemetry to a backend service that builds a device capabilities table keyed by device model identifier and serves scores via a cache. Feature gates use threshold comparisons against the continuous score to enable, disable, or adapt compute-intensive features. The score may complement other classification signals and may be generalized to other processors such as GPUs or accelerators and to other platforms including wearables, embedded devices, and edge computing systems.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Anonymous, "Aggregate CPU Core Frequency Scoring for Instant Mobile Device Performance Classification", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10608