Watch: AM4 vs AM5 vs Threadripper: Does CPU Performance Matter for Local AI?
Micro Center tested 10 AMD CPUs, from the Ryzen 5 3600XT to the Threadripper PRO 9965WX, to see how much your processor really matters for local AI.AI
Does your CPU actually matter when you're running AI locally? We know the GPU does most of the heavy lifting, but how much performance are you leaving on the table with an older processor?
The test bench was big. We tested 10 different AMD CPUs, ranging from the Ryzen 5 3600XT to the high-end Ryzen Threadripper PRO 9965WX, and paired them with an NVIDIA GeForce RTX 5090 and an RTX PRO 6000 workstation GPU. Along the way we measured AI inference performance, token generation speed, model loading times, and workloads with multiple concurrent users.
The results were surprising. Despite enormous differences in CPU architecture, core counts, and price, there was almost no measurable difference in token generation speed for many of the GPU-accelerated AI workloads.
Loading models into VRAM was a different story. The older Ryzen 5 3600XT took considerably longer than newer processors, and the video digs into why, including how PCIe bandwidth, CPU architecture, memory, and storage can affect how quickly a model gets prepared and loaded. That includes a look at a Samsung 9100 Pro SSD on an MSI A520M-A PRO motherboard to show where an AM4 platform can become a bottleneck.
The takeaway is that it depends on how you work. If you run a single model continuously, an older CPU may serve you fine. If you frequently switch between AI models, you may benefit more from a newer CPU platform.
Want to explore the hardware? Check out the AMD Ryzen Threadripper PRO 9965WX, the PNY NVIDIA RTX PRO 6000 Blackwell, the AMD Ryzen AI Halo Developer Platform, and the MSI A520M-A PRO motherboard.
More from MC News
- Run AI Locally: The Best LLMs for 8GB, 16GB, 32GB Memory and Beyond
- How to Use Local AI With Your Own Files: A Beginner's Guide to RAG
- Reading vs. Answering: The Two Hardware Bottlenecks Behind Local AI Performance
- How to Use a Local LLM in Visual Studio Code
- Hands-On with the AMD Ryzen AI Halo
- Quantization Explained: Why the Same LLM Gives Better Results on High-End Hardware
- Why Context Length, Not AI Model Size, Is Your Biggest VRAM Problem
- How to Use LM Studio to turn your PC into an AI Server
- Why VRAM and Memory Bandwidth are Key for Powering Local AI
- Keyboard 101: Intro to Computer Keyboards
- Fix It Yourself: Talking to iFixit on Why Repairable Tech Matters
- The Local LLM Your Laptop's Been Waiting For
- Gemma vs Qwen: Which Local AI is Best?
