In the age of AI, the need for GPUs such as the H100 has actually increased tremendously, making it tough for an average customer to get one. Nevertheless, a Reddit user has actually created a special service, converting the AMD Zen 2-based Ryzen 5 4600G “Renoir” APU into a 16 GB GPU and then utilizing it for AI work on Linux. You understand, as they state, contemporary issues require modern-day options.
Your Old AMD Ryzen APUs Provide Decent Performance in AI Workloads After Conversion Into A GPU
Before we enter into how the APU was converted, let’s take a quick recap of the AMD Ryzen 5 4600G. The Ryzen 5 4600G was understood to be one of the very best APUs in the market after ultimately getting changed by its Cezanne equivalent. It featured a 6C/12T setup with the Radeon Vega iGPU with 7 CUs (Compute Units). To explain how the 16 GB VRAM mark was achieved, it is vital to note that APUs support “Shared Memory”, in which you can allocate 50% of your RAM’s capacity to your APU. In this case, the Reddit user had 32 GB DDR4 memory onboard, offering half of it to the processor.
The next huge hurdle is really running AI workloads on a Ryzen APU. If you had a desktop GPU, you could utilize AMD’s ROCm (Radeon Open Compute) platform to run AI applications on Linux. However, when it comes to an iGPU, third-party packages enable ROCm to operate on APUs, which was also used here. Utilizing ROCm fixes the majority of your problems because now you can run every kind of AI application varying from Tensorflow to PyTorch.
In a comprehensive video, the Reddit user showcased his fascinating experiment, declaring that the Ryzen 5 4600G might manage all sorts of AI workloads. However, he only revealed testing of Stable Diffusion, and to our surprise, the APU successfully created a 512×512 image in around one minute and 50 seconds. This is a decent turning point for the APU, and we believe it might prove to be a breakthrough if used properly.
While you can’t compare the Ryzen 5 4600G to an AI-meant GPU, such as the H100, this experiment shows the untapped potential in Ryzen APUs. Picture if instead of the Ryzen 5 4600G, the latest Phoenix APUs were being utilized here, which are much remarkable in efficiency. Nevertheless, offering such AI performance for $95 is undoubtedly something to appreciate, and we do hope that it supplies the “typical consumer” a method to meet his AI requirements.
News Source: Tom’s Hardware