amd

Safe, reversible Linux CPU performance profiles across CPUFreq, intel_pstate, amd-pstate, cpuidle, thermal, hwmon, and future Intel/AMD/ARM backends.

One sentence in, trained computer vision model out, zero human labels. An autonomous agent swarm (FLUX.2 - SAM 3 - Gemma 4) running entirely on one AM...

llama.cpp HIP vs Vulkan benchmark on AMD Strix Halo (gfx1151) with ROCm 7.2.2 — Ubuntu 26.04

🚀 Run 120B AI Models on Spark 2026 - vLLM API & Coding Assistant

蓝桥杯 FPGA 设计与开发（AMD/Xilinx 平台）国一工程模板与零基础教学资源库，涵盖完整驱动库、真题解析、模拟赛代码和入门资料

⚡ Tu estudio multimedia con IA, entero en tu máquina — reescalado, restauración de audio, karaoke, transcripción, doblaje, generación de imagen/video...

Vulkan compute backend for PyTorch — runs on any GPU. PrivateUse1 dispatch, SPIR-V shaders, zero ROCm/CUDA dependency.

Community-curated guides, benchmarks, and compatibility notes for running LLMs on AMD hardware with ROCm, Vulkan, llama.cpp, vLLM, Ollama, and PyTorch...

Tools for AI/ML Workload Analysis and Characterization

Install AMD consumer Ryzen chipset, Radeon graphics, Ryzen AI NPU (XDNA), and the Microsoft inbox Bluetooth PAN driver on Windows Server 2016/2019/202...

Connect local tools, databases, and LLMs through a desktop interface for Windows, Linux, and macOS.

Heterogeneous Compute Cascade (HCC) — distributed 400B-parameter MoE inference across dual AMD Ryzen AI MAX+ 395 'Strix Halo' workstations via USB4.

Detect, validate, enable and assess the AMD XDNA NPU on Linux. Dependency-free. Gives an honest, machine-specific LLM-feasibility verdict for XDNA 1 (...

Run PyTorch natively on Windows 11 with an AMD RX 9070 XT (RDNA4 / gfx1201) on stable ROCm 7.2.1 — no WSL2, no Linux, no ZLUDA. Exact pinned wheel URL...

Voxen: Assembly-Aware CAD Generation System Powered by Fine-Tuned Qwen3-8B on AMD MI300X

Single-file bash system dashboard for Linux workstations. Hardware telemetry, containers, process management, security audit -> zero cloud depe...

HIPIFY-first AI agent for CUDA2ROCm migration, compiled and test-verified on AMD MI300X with Gemma 4.

Легковесная идиоматичная системная утилита на Rust для тонкого управления питанием, лимитами и частотами процессоров AMD Ryzen (архитектура **Zen 4 /...

Reproducible TRELLIS.2 INT8 ConvRot patches for ComfyUI on AMD ROCm gfx1100

TUI PCI register viewer — scan GPU devices, read config space, mmap BAR registers (NVIDIA/AMD/Intel)

[BETA] CLI utility for reading, dumping, comparing and analyzing AMD Polaris (RX 400/500) VBIOS ROMs. Parses real AtomBIOS structures - and outputs pl...

LingChat 内置 IndexTTS 的 Windows AMD ROCm 运行时与官方模型安装器

Comparative benchmark of 4 LLM inference engines (HuggingFace, vLLM, AWQ, Speculative Decoding) on AMD GPU. Streamlit dashboard with real-time GPU tel...

Lightweight terminal dashboards for monitoring ComfyUI on Ubuntu with AMD ROCm or NVIDIA CUDA GPUs. Tracks VRAM, system RAM, ComfyUI RSS and swap, mem...

A lightweight Windows batch script that disables optional AMD background services, scheduled tasks, startup entries, and processes, then verifies thei...

PowerShell tool for reading, enabling, or disabling AMD Core Performance Boost in BIOS on supported HP commercial PCs using HPCMSL. Includes real-worl...

DaVinci Resolve Linux Container

ComfyUI custom node: text-to-image via stable-diffusion.cpp (sd-cli.exe, Vulkan backend). GGUF models: Z-Image / Krea-2 / SD1.5 / SDXL / FLUX / SD3.5

AMD Renoir ACP microphone kext with DMA continuity fixes and configurable hardware profiles

See exactly where your LM Studio model weights & KV-cache live — GPU (VRAM) or system RAM. Real per-process measurement, GGUF-aware KV math (GQA/MLA/S...

Windows gaming performance profiler and CPU scheduling optimizer with per-game profiles, frametime analysis and benchmarking.

Get internal audio (TAS2783 speakers + RT712 mic on AMD SoundWire) working on the HP OmniBook X Flip 14 (board 8EA1) under Linux

Run a model too big for one GPU in LM Studio across two PCs over LAN (distributed inference via llama.cpp RPC): two DLLs + one env var GGML_RPC_SERVER...

Ternary Bonsai-27B (1.71bpw) on a 12GB AMD RDNA2 card via Vulkan — no ROCm. First published RDNA2 numbers, one-script install, full receipts.

Qwen3.8-27B EXL3 on AMD gfx1201: RX 9070 XT (16 GB, tested) and Radeon AI PRO R9700 (32 GB). vLLM plugin with custom kernels: ~80 tok/s with MTP specu...

Analyse any text or ebook on 82 calibrated dimensions (genre, mood, themes, style) with the Laya model — ~1 s per page on Strix Halo with ROCm 10. Mod...

Documentation for MVGAL, the Multi-Vendor GPU Aggregation Layer for Linux

Lightweight filter program that allows you to create, save and load custom filters

DLSS 5 on AMD — even in games that never had DirectX 12. Built and first tested on Radeon RX 7600 with GTA San Andreas.

Self-hosted AI image and video generation for AMD Ryzen AI APUs on Linux: Vulkan (Mesa RADV) via stable-diffusion.cpp, no ROCm. Web UI with queue, mod...

Run macOS 13 Ventura on WSL2 nested KVM with Reims vGPU + Dozen (Metal hardware accel, noVNC)

System monitor that reports real GPU memory on unified-memory AMD parts (Strix Halo / Ryzen AI Max), where other tools show only the firmware carve-ou...

Experimental Linux/Vulkan video player adapting AMD FSR 4.1-style temporal reconstruction to ordinary decoded video — no frame interpolation.

C++/HIP Qwen3.8-27B on 3x MI50 32GB (gfx906): 69.17 native / 110.66 committed MTP tok/s; qualified through 262K.

openPangu Flash92 — C++/CUDA & C++/HIP ROCm Engines. Fully resident, Blackwell tensor cores, no Python. Powered by openPangu.

YMM4向けAMD Radeon AMF H.264/HEVCハードウェア動画出力プラグイン

GPU tuning for NVIDIA and AMD from one executable. Clocks, voltage, V/F curve, fan control, live monitoring, and a CLI.

Deep Vulkan vs HIP/ROCm auto-benchmarker for AMD local LLMs (Ollama + LM Studio). Warm-up, median of 3 runs, VRAM unload, telemetry.

Run CUDA and HIP programs unmodified on a CPU. No GPU needed: simulated NVIDIA and AMD GPUs for development, teaching and CI.

262K-context Qwen3.8-27B on a single RX 7900 XTX: llama.cpp ROCm, IQ3_S+MTP n=3, KV q8_0/q5_1, -ub 256, 272 W. 19-27 tok/s at 240K fill, 60/60 recall.