KlyveroMulti Worker Benchmark
Compare browser worker workload throughput.
Klyvero Benchmark v1 · Browser performance only
How does your browser handle everyday tasks?
Working on several tasks. Each result explains what was measured and how to read it.
About 15 seconds · Keep this tab open and close busy apps first. Your device may get warm and use more battery. You can stop at any time.
Optional. Keep this the same during the test. If you plug in or unplug, stop and start again.
Ready when you are
How should I use these results?
Start with the task, not the biggest number. More tasks per second means more work completed; fewer milliseconds means less waiting. 1,000 milliseconds equals one second. Different checks use different tasks and units, so their numbers cannot be compared with one another.
To compare repeated runs, use the same test version, browser, settings and power conditions. Technical details show the method, worker count and source hash needed to check whether runs match.
There is no overall score or device ranking yet. We do not have a reviewed reference dataset to say how your device compares with others. Score calibration pending. These are approximate browser measurements, not a hardware health certificate.
Bound concurrency for power/responsiveness; logical processor hint is not physical core inventory
How to use this test
- Run the bounded parallel worker workload and note the effective worker count.
- Choose a workload and explicitly start the versioned benchmark. You can stop at any time.
- Review seven raw samples and dispersion. Score calibration is pending; native benchmark equivalence is not claimed.
How it works & what results mean
How the test works
Compare browser worker workload throughput. It uses Worker; performance.now; optional hardwareConcurrency hint. The workspace shows observations from this browser or asks for your own confirmation.
Understanding your result
Dispatch, barriers and scheduling are included. Compare only equal source hashes and worker-count classes; logical processor hints are not physical cores. Bound concurrency for power/responsiveness; logical processor hint is not physical core inventory Missing data, blocked permission and an operation that could not run are shown separately. None establishes a hardware fault.
Keep checking
Related tests
About this test & common questions
how to benchmark parallel browser workers?
This multi worker benchmark runs parallel JavaScript work through Web Workers. Inspect how much of this workload the browser completes under the selected conditions.
does a multi worker test measure every CPU core?
The browser and operating system schedule workers; worker count does not prove that every physical CPU core was fully exercised.