Lukas Vogel’s SQLDoom is his second SQL-based Doom experiment after DOOMQL, which appeared in 2025. DOOMQL used raycasting and looked closer to Wolfenstein or 3D Monster Maze. SQLDoom uses BSP-tree traversal to produce visuals closer to the original game.

The new port moves Doom’s game logic and renderer into SQL. CedarDB must return either a table or a bitmap containing exact RGB values for every pixel. SQL also runs the game loop. Python handles timing and keyboard input, then displays the bitmap. The project shares no code with an existing Doom port.

Vogel said the full port proved feasible and performed better than the earlier DOOMQL approach. He later added deathmatch. The game runs on a laptop with an AMD Ryzen 7 7840U, and online matches use servers in the EU and US. The Register found the server performance sluggish, with the servers a likely cause.

CedarDB compiles complex queries into machine code. SQLDoom serves as a demonstration of that capability, while also putting a playable Doom inside a database.