What is RDNA 4?
AMD's 2024 GPU architecture. Notable IPC improvement over RDNA 3, improved AI inference throughput, paired with GDDR6 in the RX 9060 XT series.
Full Explanation
RDNA 4 is AMD's 2024 GPU microarchitecture, powering the RX 9070 and RX 9060 XT consumer cards. Key improvements over RDNA 3 include a redesigned compute unit with higher IPC (instructions per clock), improved AI matrix acceleration throughput, and a refined memory subsystem. The RX 9060 XT ships with 16 GB GDDR6 — more VRAM than the competing RTX 5070 at a similar price. ROCm support for RDNA 4 was available at launch, with official Ollama support following within months.
Why It Matters for Local AI
RDNA 4 is AMD's most competitive AI inference offering to date. The 16 GB GDDR6 capacity advantage over the RTX 5070 makes it appealing for users who prioritize running larger models over raw token throughput. Linux users benefit most; Windows ROCm support is functional but requires manual configuration.
Hardware Relevant to RDNA 4
gpu · Check Price on Amazon · 16 GB VRAM · 288 GB/s
Related Terms
ROCm→
AMD's open-source GPU compute platform — AMD's answer to NVIDIA CUDA. Required for GPU-accelerated AI on AMD cards. Mature on Linux; less reliable on Windows.
GDDR6→
Previous-generation GPU memory. Lower bandwidth than GDDR7, but paired with larger capacities (e.g., 16GB RX 9060 XT) can offer better model headroom despite lower token speed.
VRAM→
Video RAM — dedicated memory on a GPU. Determines the maximum model size you can run with full GPU acceleration. Once a model exceeds VRAM, it spills to system RAM over the slow PCIe bus.
CUDA→
NVIDIA's proprietary parallel computing platform. Industry standard for AI/ML. Nearly every AI framework (PyTorch, Ollama, ComfyUI) supports CUDA natively and first.