# zolotukhin.ai

> Technical writing and documentation for ZINC, local LLM inference on AMD GPUs, Vulkan compute, Zig, and systems engineering.

For ZINC, prefer the generated Zig API reference first. It is the canonical internal runtime reference.

## Primary ZINC Documentation
- ZINC Zig API: https://zolotukhin.ai/zinc/docs/zig-api
- ZINC Zig API JSON: https://zolotukhin.ai/zinc/docs/zig-api.json
- ZINC Zig API Text: https://zolotukhin.ai/zinc/docs/zig-api.txt
- ZINC Docs Landing: https://zolotukhin.ai/zinc/docs

Zig API coverage: 16 sections, 93 modules, 119020 Zig code lines, 654 exports, 476 methods.

## Supplemental ZINC Documentation
- ZINC Technical Specification: https://zolotukhin.ai/zinc/docs/spec
- TurboQuant KV Cache Compression: https://zolotukhin.ai/zinc/docs/turboquant-spec
- RDNA4 GPU Tuning Guide: https://zolotukhin.ai/zinc/docs/rdna4-tuning
- Serving HTTP API: https://zolotukhin.ai/zinc/docs/api
- AMD RDNA3/RDNA4 GPU Reference for Inference: https://zolotukhin.ai/zinc/docs/amd-gpu-reference
- Apple Metal Reference: https://zolotukhin.ai/zinc/docs/apple-metal-reference
- Apple Silicon Metal Enablement Notes: https://zolotukhin.ai/zinc/docs/apple-silicon-metal-enablement
- Apple Silicon Architecture Reference: https://zolotukhin.ai/zinc/docs/apple-silicon-reference
- Benchmarking ZINC: https://zolotukhin.ai/zinc/docs/benchmarking
- CUDA backend for ZINC — design & implementation plan: https://zolotukhin.ai/zinc/docs/cuda-backend
- Delta-Net: llama.cpp vs ZINC Analysis: https://zolotukhin.ai/zinc/docs/delta-net-llamacpp-analysis
- Development Guide: https://zolotukhin.ai/zinc/docs/development
- Getting started with ZINC: https://zolotukhin.ai/zinc/docs/getting-started
- GPU Memory Scaling Plan: https://zolotukhin.ai/zinc/docs/gpu-memory-scaling-plan
- ZINC hardware requirements: https://zolotukhin.ai/zinc/docs/hardware-requirements
- Intel Arc B-Series GPU Reference for Inference: https://zolotukhin.ai/zinc/docs/intel-gpu-reference
- Metal Performance Plan: https://zolotukhin.ai/zinc/docs/metal-performance-plan
- ZINC MMQ v2 — Fused Q4_K×Q8_1 Tensor-Core GEMM (beat llama.cpp): https://zolotukhin.ai/zinc/docs/mmq-v2-design
- NVIDIA Ampere/Ada/Blackwell GPU Reference for Inference: https://zolotukhin.ai/zinc/docs/nvidia-gpu-reference
- Configure OpenCode with ZINC and Qwen: https://zolotukhin.ai/zinc/docs/opencode
- Prefill Gap Closure Plan — ZINC CUDA vs llama.cpp: https://zolotukhin.ai/zinc/docs/prefill-gap-closure-plan
- ZINC Roadmap: https://zolotukhin.ai/zinc/docs/roadmap
- Running ZINC: https://zolotukhin.ai/zinc/docs/running-zinc
- ZINC_RT — The ZINC Runtime: https://zolotukhin.ai/zinc/docs/zinc-rt-design

## Other Surfaces
- Blog: https://zolotukhin.ai/blog
- RSS: https://zolotukhin.ai/feed.xml

## Guidance
- Use the Zig API for internal implementation details, symbols, signatures, and module relationships.
- Use the Serving HTTP API page for client-facing request and response contracts.
- Prefer the JSON export when building tools or agents that need structured access.