Developer Tools●●Solid
HeatSpectra: Realtime 3D Surface Heat Simulation
Meshless Voronoi discretization running realtime on Vulkan is genuinely clever.
WizardryNiche Gem
tsun3doku
602mo ago
Back to browse
GitHub Repository
Experimental Godot 4.7 project running Gemma 4 inference entirely in GDScript and Vulkan compute shaders
1 starsGDScript
Godot running Gemma 4 inference in GDScript and Vulkan compute shaders
by steamspaghetti·Jul 12, 2026·1 point·0 comments
AI Analysis
●●●BangerWizardryBig BrainRabbit Hole
Running LLM inference in pure GDScript and Vulkan shaders inside a game engine.
Strengths
- •Implements transformer architecture using only Godot's native compute shaders.
- •Handles GGUF parsing and quantized tensor reads without external native extensions.
- •Demonstrates full inference loop including tokenizer and sampling in game script.
Weaknesses
- •Inference speed is an order of magnitude slower than optimized C++ stacks.
- •Hardcoded to a single specific model artifact with no runtime selection.
Category
Target Audience
Game developers and graphics programmers
Similar To
llama.cpp · MLC LLM
Post Description
Experimental Godot 4.7 project running Gemma 4 inference entirely in GDScript and Vulkan compute shaders.
Similar Projects
AI/ML●●●Banger
Valkyr LM Inference with Realtime Guarantees
Pure Vulkan compute enables LLMs inside game loops without CUDA lock-in.
WizardryNiche Gem
quatonion
302mo ago
Gaming●●Solid
10k moving agents in Godot using GDScript
Shared field approach moves 10k agents at 5ms where per-agent queries collapse at 500.
Big BrainNiche Gem
Vav-Labs
2027d ago
AI/ML●●Solid
Gemma 3 inference in pure C++ with Metal acceleration
Pure C++23 Metal inference when llama.cpp already dominates this space.
WizardryNiche Gem
ybubnov
628d ago
Infrastructure●●Solid
Go LLM inference with a Vulkan GPU back end that beats Ollama's CUDA
28% faster Vulkan-to-CUDA on Qwen, but llm.c and llama.cpp already own inference.
WizardryBig BrainNiche Gem
computerex
104mo ago
AI/ML●●Solid
An experimental tensor DSL/compiler with explicit scheduling
Three-concept scheduling model targeting FlashAttention, but generated code isn't performant yet.
Big BrainBold Bet
loganboyd
511mo ago