![[PS4] Review of the Ever 17/Never 7 Double Pack](https://www.streamgamer.net/wp-content/uploads/2025/02/PS4-Review-of-the-Ever-17Never-7-Double-Pack-360x180.jpg "[PS4] Review of the Ever 17/Never 7 Double Pack")
AMD’s Radeon RX 7900 XTX edges out NVIDIA’s GeForce RTX 4090 in performance when running the DeepSeek R1 AI model, showcasing impressive results in inference benchmarks.
## AMD’s Swift Response to DeepSeek’s R1 LLM Models Puts Them Ahead
DeepSeek’s latest AI model has made waves in the industry, leaving many curious about the computing muscle required to train it. In an interesting turn, it turns out that everyday users can achieve commendable performance using AMD’s Radeon RX 7900 XTX GPU from the “RDNA 3” series. AMD took the opportunity to share inference benchmark comparisons between their top-of-the-line RX 7000 series GPU and NVIDIA’s offering, demonstrating superior performance across various models.
DeepSeek performing very well on @AMDRadeon 7900 XTX. Learn how to run on Radeon GPUs and Ryzen AI APUs here: pic.twitter.com/5OKEkyJjh3
— David McAfee (@McAfeeDavid_AMD) January 29, 2025
Leveraging consumer GPUs for AI tasks has proven advantageous for many, largely due to the solid price-to-performance value these offer compared to standard AI accelerators. Additionally, running models locally with these GPUs helps ensure privacy—a significant concern surrounding DeepSeek’s AI models. Fortunately, AMD has rolled out a detailed guide on operating DeepSeek R1 distillations on their GPUs, which goes as follows:
1. Ensure your system is using the 25.1.1 Optional or higher Adrenalin driver.
2. Download LM Studio 0.3.8 or later from lmstudio.ai/ryzenai.
3. Install LM Studio and skip past the onboarding screen.
4. Navigate to the discover tab.
5. Select your preferred DeepSeek R1 Distill. For blazing speed, start with smaller distills like the Qwen 1.5B, though larger distills will offer better reasoning capabilities—each very robust.
6. On the right, select “Q4 K M” quantization and click “Download.”
7. Once downloaded, go to the chat tab, select the DeepSeek R1 distill from the dropdown menu, and check “manually select parameters.”
8. In the GPU offload layers, slide all the way to maximum.
9. Click model load.
10. You’re all set to interact with a reasoning model operating entirely on your local AMD hardware!
Should these steps prove tricky, AMD has produced a YouTube tutorial detailing each part of the process. It’s worth a look if you aim to run DeepSeek’s LLMs on your own AMD setup while keeping your data protected. Anticipation is brewing for the next wave of GPUs from NVIDIA and AMD, with expectations for significant boosts in inference capabilities, thanks to the onboard AI engines designed to handle these demanding tasks more effectively.
