Frame Budget Lab
Make frame-time tradeoffs visible before you optimize. Tune CPU and GPU costs, find your bottleneck, compare refresh-rate targets, save your scenario locally and export a report. No account, tracking or external API.
01 / Workload model
Scenario stays in this browser only.02 / Live performance readout
Calculate
03 / Target feasibility
| Target | Frame budget | Result |
|---|
This is an educational throughput model. It assumes overlap between total CPU and total GPU work; real engines also face synchronization, jitter, memory pressure and device-specific scheduling. Validate results with frame traces on your target hardware.
How to use these numbers
A 60 FPS refresh target gives roughly 16.67 ms per frame. In this model, CPU tasks add along the CPU path, while GPU tasks add along the GPU path. Effective frame time is the longer path plus a configurable synchronization overhead. A negative headroom means that workload cannot maintain the selected target under these assumptions. Trim the most expensive path first.
Example: if a CPU path takes 12 ms and the GPU path takes 20 ms, speeding the CPU from 12 to 8 ms does not change a 20 ms GPU bottleneck. Reduce GPU passes or resolution instead. Aim for a margin to absorb frame-time spikes; average values are not a guarantee of smooth frame pacing.