Quick Answer
Our CPU benchmark test tool is a browser utility at /run/ that executes configurable workloads, generates performance reports with throughput and stability metrics, detects logical core count, and exports JSON for result comparison.
Formula
Report = { singleCoreScore, multiCoreScore, throughput, stabilityPct, duration, intensity, threadMode, timestamp }
Introduction
You do not need to install software to validate processor performance. Our tool runs entirely in the browser, stores nothing on a server, and produces shareable JSON reports for longitudinal comparison.
This guide covers benchmark execution controls, score generation logic, and how to use exported data for result comparison across passes, power profiles, and hardware changes.
What does the benchmark test tool provide?
The tool at /run/ offers an intensity slider (1-40), duration presets from 30 seconds to 5 minutes, workload type selection, and thread mode options including Auto, Single-Core, Multi-Core, and Extreme.
During benchmark execution, live metrics update throughput, normalized scores, estimated CPU load, and stability percentage. Hardware detection displays logical core count and browser environment info.
When the run completes, a performance report modal summarizes all metrics. JSON export captures every field for result comparison in spreadsheets or version control.
Score generation mechanics
The engine tracks operations completed per second during each phase, then maps throughput to normalized 0-100 scores for single-thread and multi-thread performance separately.
Stability percentage samples throughput at regular intervals. Consistent samples produce high stability; declining samples indicate throttling or interference during benchmark execution.
Displayed Score = normalize(opsPerSecond) per phase
- Integer workload: prime sieve style loops
- Float workload: matrix multiplication kernels
- Mixed workload: alternates integer and float stress
- Crypto workload: hashing-style compute patterns
Step-by-step: executing a benchmark and reading the report
Complete workflow from opening the page to archiving results.
-
Open /run/
Load the page in a modern Chromium, Firefox, or Safari browser.
-
Verify hardware detection
Confirm logical core count matches your system. Incorrect counts may indicate browser limits.
-
Configure test parameters
Start at intensity 10-15 for laptops, 20+ for desktops. Use 120-300 second duration for stability analysis.
-
Execute benchmark
Click Start Test. Monitor live throughput and stability during the run.
-
Review performance report
Open the results modal. Note single-thread, multi-thread, throughput, and stability fields.
-
Export for comparison
Download JSON. Label files with date, power mode, and test settings for valid result comparison.
Example: tracking a BIOS update impact
A user exports a baseline report: single-thread 76, multi-thread 91, stability 95%. After a BIOS update, a second export reads 76, 89, 88%.
Multi-thread dropped 2% but stability fell sharply. Investigation reveals the update changed fan curves, causing earlier thermal throttling under sustained load.
Result comparison validated a regression that headline scores alone would have missed. The user rolled back the BIOS and restored 94% stability.
FAQ
- Are performance reports uploaded?
- No. Reports generate locally. Export JSON manually if you want to archive or share results.
- Which duration is best for stability analysis?
- Use at least 120 seconds. Five-minute runs best expose thermal throttling and power limit behavior.
- Can I compare JSON exports across devices?
- Yes, if settings match. Note that different browsers and core counts affect absolute scores. Compare trends and ratios, not just raw numbers.
Conclusion
Our CPU benchmark test tool delivers instant performance reports with hardware detection and JSON export for result comparison.
Use consistent settings and label exports with environmental metadata for valid longitudinal analysis.
Open the Benchmark Tool