Validation Methodology
How we measure, validate, and interpret CPU performance
Run CPU Benchmark TestWhat we measure
Our CPU benchmark test captures processor throughput, single-thread responsiveness, multi-thread scaling, and score stability under sustained compute load. Workloads include integer sieving, floating-point matrix math, mixed kernels, and optional crypto-style hashing.
Single-thread phases run on the main browser thread. Multi-thread phases distribute parallel work through Web Workers mapped to logical processors. Extreme mode combines both for maximum stress validation.
How scores support workload analysis
Single-thread scores correlate with application responsiveness and lightly threaded workloads. Multi-thread scores correlate with parallel pipelines like rendering, compilation, and virtualization.
Stability percentage reveals whether performance holds steady or degrades from thermal throttling, power limits, and background interference during the test window.
Validation principles
We recommend multiple passes with identical settings before drawing conclusions. Document power mode, cooling state, and environmental factors alongside every export.
Browser-based testing reflects JavaScript engine performance and Web Worker scheduling in addition to raw hardware capability. Use results for relative validation on your hardware; cross-tool score equivalence requires separate native benchmarks.