STAGE 1 OF 3
CPU Stress Test
Stage 1 subjects the CPU to 60 seconds of sustained, parallel compute workload across all available logical cores. Unlike synthetic burst benchmarks that measure peak clock speed in isolation, this test measures sustained throughput — the performance a CPU can maintain when thermal and power limits are actually engaged. The drop from peak to valley is the primary diagnostic signal.
What is Measured
Decay % is the key forensic signal. A decay of 3% indicates the CPU maintains near-burst performance throughout the test. A decay of 30%+ indicates a large sustained-throughput drop that can be consistent with thermal limits, power limits, or competing system activity. This browser test does not read temperature sensors directly.
How the Test Works
The benchmark spawns Web Workers to saturate all logical CPU cores with a synthetic compute workload (floating-point operations and integer math). Throughput (ops/sec) is sampled every 500ms throughout the 60-second window. The live chart renders in real time so you can watch the throttle curve develop.
Stage 1 runs automatically after the hold-to-start sequence on /benchmark-test. When it completes, a 1-second cooldown overlay plays before advancing to Stage 2.
Scoring Thresholds (CPU Decay)
- S grade: Decay ≤3% — excellent sustained throughput retention
- A grade: Decay ≤8% — strong sustained throughput retention
- B grade: Decay ≤15% — moderate measured throughput loss
- C grade: Decay ≤25% — substantial measured throughput loss
- D grade: Decay >25% — severe measured throughput loss; investigate and retest
Possible Causes of High CPU Decay
- Dried or improperly applied thermal paste (most common after 2–3 years)
- Dust-clogged CPU cooler heatsink fins
- Insufficient case airflow (intake/exhaust imbalance)
- Aggressive power limits in BIOS (especially laptop OEMs)
- Background processes competing for cores during the test window
Next Stage
After Stage 1 completes, the test automatically advances to Stage 2 — GPU Compute Test.