ROCm
AMD's open-source answer to CUDA — the gap that narrowed at last.
You're looking at a software record: a tool or platform in the AI stack — who maintains it and what hardware it runs on.
ROCm is AMD's open GPU compute stack: HIP (a CUDA-like language with automated porting), libraries, and framework support. Its immaturity long limited Instinct sales more than the silicon did; by ROCm 6–7, PyTorch and vLLM ran well enough for Microsoft, Meta and OpenAI deployments, and AMD develops it in the open on GitHub.
Specifications & key facts
Each figure is labeled by basis: vendor spec, vendor claim, reported, estimate, or official. "Not publicly disclosed" means exactly that — we never invent values.
Overview
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Supports
Maintained by
Maintains
Competes with
Sources
Record reviewed as of 2026-01-15 — see Methodology.