FPS Bottleneck Calculator

How it is calculated

You supply two frame rates measured for the same game and settings: what the CPU could sustain if the GPU were unlimited (cpu_fps), and what the GPU could sustain if the CPU were unlimited (gpu_fps), both in frames per second. The system runs at the slower one: effective = min(cpu_fps, gpu_fps). It is shown as a whole number from 100 FPS up and when it is already whole, otherwise with one decimal from 10 FPS and two decimals below that (47.5 stays 47.5). The bottleneck percentage is the share of the faster part that goes unused: (faster − effective) ÷ faster × 100. Utilisation of each part is effective ÷ its own fps × 100, so the limiting component always shows 100%. The tool does not name the limiting part; read it from the 100% figure. Both values must be above zero. It does not estimate the fps for you.

Common mistakes

• Entering the same in-game fps in both boxes. The tool needs two different numbers: one from a CPU-limited test (low resolution, low settings) and one from a GPU-limited test (high resolution, high settings). Equal values just show a 0.00% bottleneck.
• Reading the bottleneck percentage as extra fps you would gain. A 37.50% result means the faster part is 37.50% idle, not that swapping the slower part adds 37.50% fps.
• Treating the percentage as 'bad' on its own. Some gap is normal, and a GPU-bound system at 100% GPU use is usually how a gaming setup is meant to run.
• Mixing numbers from different games, resolutions or scenes. The CPU and GPU figures only mean something when measured on the same workload.
• Assuming a bottleneck stays fixed. Raising resolution usually moves the limit toward the GPU, lowering it toward the CPU, so rerun with your actual settings.