Three months wrong about why my 4-node AMD cluster was slow [qcjfB428y4r]
Three months on a 4-node Minisforum MS-S1 Max Strix Halo cluster, and the one assumption that was quietly killing my inference speed. Try out ChatLLM - and Abacus AI DeepAgent - 🛒 Gear Links 🛒 👀 2 400Gbps switch: 👀👀 4 400Gbps switch: 💻☕ Thunderbolt 5 external SSD: 💻☕ Favorite 15" display with magnet: 🎧⚡ Great 40Gbps T4 enclosure: 🛠️🚀 My nvme ssd: 📦🎮 My gear: 🎥 Related Videos 🎥 🏆 Skip M3 Ultra & RTX 5090 for LLMs | NEW 96GB KING - 💻 Smallest RTX Pro 6000 rig | OVERKILL - 🔧 Cheap mini runs a 70B LLM 🤯 - 🌙 RAM torture test on Mac - 🚀 FREE Local LLMs on Apple Silicon | FAST! - 🪞 REALITY vs Apple’s Memory Claims | vs RTX4090m - 📦 Set up Conda - 🤖 INSANE Machine Learning on Neural Engine - * 🛠️ Developer productivity Playlist - 🔗 AI for Coding Playlist: 📚 - — — — — — — — — — ❤️ SUBSCRIBE TO MY YOUTUBE CHANNEL 📺 Click here to subscribe: — — — — — — — — — Join this channel to get access to perks: — — — — — — — — — 📱 ALEX ON X: Donato's channel: #macmini #llm #nvidia