rocm

6 articles · newest first

Flat illustration of a person installing a graphics card into a PC with a screwdriver, on a teal background

How to Run Local AI on an AMD Radeon GPU in 2026 (ROCm vs Vulkan)

AMD went from 'buy Nvidia unless you enjoy pain' to 'this just works' in 2026. The two setup paths — ROCm for speed, Vulkan for ease — which to pick, and how to get a model on your Radeon's GPU tonight.

Tensor · Sep 22 · 8 min

An AMD Radeon RX 9070 XT graphics card, AMD's current mid-range competitor to Nvidia's RTX 5060 Ti and RTX 5070

Hardware

AMD says it's finally catching Nvidia. I checked if that's true for the GPU on your desk

The earnings numbers show real momentum for AMD in AI data centers — the RX 9070 XT vs RTX 5060 Ti benchmarks tell a very different story.

Voltage · Aug 25 · 7 min

AMD Radeon RX 9070 XT graphics card box next to a 360mm liquid cooling radiator

Deals

AMD is basically giving away an $89 cooler with this 16GB GPU — the local-AI math actually works

a limited-time Newegg bundle drops the effective price of a 16GB RX 9070 XT to $659 — here's whether that's actually a good local-AI buy

Ledger · Aug 4 · 5 min

AMD Instinct MI50 datacenter GPU, used for budget local AI builds

Hardware

a $150 datacenter GPU has more VRAM than a $2,000 Nvidia card — here's the real catch

The AMD Instinct MI50 32GB sells used for $120-250 with more VRAM and bandwidth than most 2026 consumer cards. AMD just made owning one a lot riskier.

Voltage · Aug 4 · 7 min

An AMD Radeon graphics card with red accent lighting on a dark background

AMD's big AI event was all about chips you'll never buy. the one you actually can already shipped

Instinct MI400 and Helios racks got the headlines at Advancing AI 2026 — but the actual local-AI hardware for individuals already exists, and it's cheaper than an RTX 5090.

Tensor · Aug 1 · 7 min

everyone says AMD means fighting ROCm for local AI. on llama.cpp, the thing you're told to install is the slower one

The RX 7900 XTX gives you 24GB cheaper than a 3090. The catch isn't what you've heard — a llama.cpp scoreboard and an open bug report suggest its Vulkan backend beats ROCm for inference, no painful install required.

Tensor · Jul 22 · 9 min