TopicsTopic
amd
6 posts from the bench tagged amd, newest first.
Tutorial
OpenClaw Part 4: Local Model Shootout on Strix Halo
Benchmark local LLMs on AMD Strix Halo's 128GB unified memory. Compare speed and answer quality across models to pick the best one for your OpenClaw agent.
local-ai · amd
~60 minDifficulty: intermediate
Tutorial
OpenClaw on AMD Strix Halo: Set Up a Self-Hosted AI Agent with 128GB Unified Memory
Set up a self-hosted OpenClaw AI agent on AMD Strix Halo: upgrade the kernel, install ROCm 7.2, and configure 128GB of unified memory for running local models.
local-ai · amd · homelab
~45 minDifficulty: intermediate
Tutorial
ComfyUI + ROCm 7.1 + PyTorch 2.8 + Strix Halo
Install ComfyUI with ROCm 7.1 and PyTorch 2.8 on an AMD Strix Halo system running Ubuntu 24.04 for Stable Diffusion image generation without an NVIDIA GPU.
local-ai · comfyui · amd
~3 minDifficulty: intermediate
Tutorial
Strix Halo + ROCm 7.1 + Ubuntu 24.04
Install ROCm 7.1 on the AMD Strix Halo (GMKtec EVO X2, AI MAX+ 395) running Ubuntu 24.04 with kernel 6.14 upgrade, AMD GPU drivers, and shared memory (TTM) configuration.
local-ai · amd · ubuntu
~6 minDifficulty: intermediate
Tutorial
Ollama for AMD Integrated GPUs in Windows
If you try and run Ollama with an integrated GPU most likely it will run on the CPU only. This guide shows you how to build a custom version of Ollama with the latest HIP SDK to enable GPU acceleration on AMD Radeon iGPUs like the 680M, 780M, and 890M.
local-ai · ollama · amd
~15 minDifficulty: advanced
Tutorial
ComfyUI with ZLUDA for AMD GPUs on Windows
ZLUDA (CUDA on AMD) is a compatibility layer that allows CUDA applications to run on AMD GPUs. This guide walks you through installing and configuring ComfyUI ZLUDA specifically optimized for AMD hardware, including the AMD Radeon 780M integrated graphics.
local-ai · comfyui · amd
~24 minDifficulty: advanced