rdna2

The rDNA 2 dGPU support patch kext.

Windows-only version of ComfyUI which uses AMD's official ROCm and PyTorch libraries to get better performance with AMD GPUs. [auto-installation and popular performance enhancing packages like triton * sage-attention * flash-attention * bitsandbytes included ]

Utilities for accessing AMD's Machine-Readable GPU ISA Specifications.

HIP/ROCm fork optimized for AMD RDNA2 (gfx1030) with PrismML Q1_0_G128 1-bit quant support, RotorQuant, TurboQuant, EAGLE3 and P-EAGLE speculative decoding, and full Wave32 kernel optimizations.

KytyPS5-based PS5 emulator with extra compatibility/HLE work. Experimental crash/boot-reach improvements — not a playability claim.

Steampunk gaming Linux distro for the AMD BC-250 (BC250, gfx1013 / Cyan Skillfish). Debian + KDE Plasma, live 40-CU control and overclocking, Steam/EmuDeck, on-device AI.

Unofficial community guide for the AMD BC-250 (Cyan Skillfish) mining board — repurposed as a budget Linux gaming/desktop PC. Hardware specs, BIOS, cooling, OS setup, GPU governor, game benchmarks, troubleshooting & more.

AMD ROCm (gfx1030) inference fork with RotorQuant/TurboQuant KV compression, PHANTOM-X zero-copy draft speculation, EAGLE3 speculative decoding, 12 RDNA2 crash fixes, and PrismML Bonsai Q1_0_G128 1-bit GGUF support.

The newcomer's bible for the AMD BC-250 — bilingual EN/RU handbook, zero-to-gaming golden path, FAQ/troubleshooting, and curated awesome resources distilled from 125k+ community messages.

Fork of librocdxg that adds ROCm support for RDNA2 GPUs under WSL (which are officially unsupported).

Exact FlashAttention for AMD RDNA2 (gfx1030 / RX 6800 XT), written in HIP. Drop-in replacement for flash-attn: 19.8 TFLOP/s forward, decode at 99% of HBM bandwidth.

Custom ComfyUI node to load .latent files directly from disk, enabling VRAM-efficient high-resolution workflows on limited GPUs (especially AMD RDNA2 / ROCm)

The newcomer's bible for the ASRock AMD BC-250 — a PS5-derived APU board turned into a budget Linux gaming & AI mini PC. 15 languages.

The bridge between your web-based Slime pet and local AI. Power your Tamagotchi with Llama 3 & OpenClaw using .NET 9 and your local GPU. 🧪⚔️

Local autonomous coding agent stack for AMD RX 6700 XT (RDNA2): OpenCode + LM Studio Vulkan + Gemma 4 E4B Q6_K + oh-my-opencode. Bypasses Ollama Vulkan + Gemma 4 upstream bug.

Custom SPIR-V kernel factory for PHANTOM speculative decoding — LLVM IR to GPU (SPIR-V/HIP) and CPU (native x86) cross-target compilation, RDNA2/gfx1030 optimized pre-compiled kernels, dynamic kernel swapping, zero-JIT inference pipeline

Local AI on AMD (RDNA2/ZLUDA) on Windows: Stable Diffusion XL, Z-Image, and Qwen3-TTS voice cloning — battle-tested setup, the cuDNN fix, the arbitrary-torch-on-ZLUDA recipe, and honest benchmarks.

Tiny Chrome extension that disables backdrop-filter blur on chromium based browsers (Chrome, Brave..) on every site video element to fix blurry playback for AMD GPUs that use hardware acceleration via D3D11.

Experimental AMD BC-250 GPU CU/WGP and CPU core unlocking suite with hardware validation, bilingual GUI, SteamOS support and reversible persistence.

ONNX Runtime acceleration via ROCm 7.1 and MIGraphX for AMD RDNA2 GPUs (RX 6600M).