Case study 01 / 07 · ● Live · 2025
FPS Calculator
Predicts real-world gaming FPS for any CPU + GPU combo across 50+ titles.
Backend on PostgreSQL with migrations, a localized React frontend, Docker deployment, and automated pytest runs on GitHub Actions.
- FastAPI
- Python
- PostgreSQL
- React 19
- Tailwind CSS
- Docker
- Nginx
50AAA titles
5Languages
3Resolutions
3Upscalers modeled
01 / The problem
A yes/no “will my PC run this” check isn’t very useful, because the CPU/GPU balance changes by game, resolution, and upscaler.
02 / The approach
A calibrated frame-time model that splits each game’s frame time between CPU and GPU work using a per-game CPU fraction, then applies resolution, preset, and DLSS/FSR/XeSS adjustments.
03 / The catch
The hardware scoring has to match its sibling site, Bottleneck Calculator, since both read the same benchmark API.
Highlights
- Per-game CPU fraction instead of a genre default
- VRAM penalty modeling for VRAM-limited cases
- Content written for AI search citation as well as Google
- Admin dashboard for game management, usage analytics, and CSV export
Like what you see?