Quick Answer
FPS testing under load pushes CPU, GPU, and thermals with high settings, dense scenes, and extended duration to reveal limits average tests miss.
Formula
Sustained FPS ≤ Burst FPS when thermal or power limits apply
Introduction
This guide is part of the FPS Test knowledge base. Use the FPS test tool on the run page for live browser measurement, then work through the sections below to interpret frame delivery quality beyond a single average number.
Short tests lie. This article covers high settings stress, CPU vs GPU bound scenes, multiplayer variability, and long-session thermal behavior so you know worst-case FPS before it ruins a ranked match or export deadline.
Load Types
High graphics settings raise shader, shadow, and texture load on the GPU.
CPU-intensive scenes with many AI, physics, or simulation ticks limit FPS regardless of GPU headroom.
GPU-intensive scenes with particles, volumetrics, and ray tracing saturate raster and RT cores.
Multiplayer performance adds network variance and higher entity counts.
Long-session testing exposes heat soak and power limit throttling.
Load testing without a stability lens misreads the problem. Combine these sessions with FPS stability test analysis when FPS falls but averages still look acceptable on paper.
- High graphics settings stress
- CPU-intensive simulation and AI scenes
- GPU-intensive particles and RT effects
- Multiplayer crowded instances
- Long-session thermal and power testing
Burst vs Sustained
First-minute FPS often exceeds tenth-minute FPS on laptops and small form factor PCs.
Use the browser tool five-minute duration plus in-game extended play for combined insight.
Resolution multiplies load nonlinearly. When stress tests fail only at native res, walk tiers using FPS test by resolution before assuming you need new hardware.
Sustained min FPS is a better upgrade trigger than burst peak FPS. Peaks sell laptops; sustained numbers reflect real ranked sessions.
Sustained FPS = min(FPS over session window)
- Increase scene complexity gradually on /run/
- Run unplugged laptop tests separately
- Monitor clock speeds if FPS drifts down
- Clean dust before blaming settings
Under-Load Protocol
Find your worst realistic case. Stress that is harder than anything you play wastes time; stress below your real scenes misses throttling.
-
Max relevant settings
Raise only settings you actually use, not pointless ultra on unused effects.
-
Pick worst scene
Dense combat, city hub, or simulation peak. Community benchmark routes help when built-in tests are too light.
-
Extend duration
Minimum five minutes after warm-up. Ten to fifteen on thin laptops if thermals are suspect.
-
Log sustained min
Compare first minute vs last minute averages. Drift implicates thermals or power limits.
-
Mitigate thermals
Improve airflow or cap FPS if throttling persists. Cleaning dust filters is free performance recovery.
-
Separate CPU vs GPU tests
Lower resolution once; if FPS barely changes, CPU or simulation bound. If FPS jumps, GPU bound.
Load Examples
Browser complexity 10, five minutes: stability score drops 15% on laptop battery.
In-game RT on in rain scene: GPU 99%, CPU 40%, FPS halved vs RT off.
MMO city hub with 80 players: CPU thread saturated, GPU idle; resolution changes ineffective.
Background video encode plus game: shared GPU causes lows unrelated to game settings until encode finishes.
- Raid encounter with full raid size
- Flight over city vs empty field
- Video export while game minimized (background load)
FAQ
- Why does FPS drop over time?
- Thermal throttling, power limits, or background processes increasing GPU use.
- CPU or GPU bound?
- If lowering resolution barely helps, likely CPU bound. If lowering resolution helps a lot, likely GPU bound.
- How hard should browser stress be?
- Increase complexity until FPS stabilizes below your target to probe headroom, but stop if device overheats.
- Are five-minute tests enough?
- Often yes for desktop towers with good airflow. Thin laptops and SFF PCs may need ten to fifteen minutes to reveal heat soak.
Conclusion
Under-load FPS testing reveals sustained limits. Combine high settings, worst scenes, and long duration with thermal awareness.
Run Load Test