To perform AI tasks on your computer such as AI image generating, you may need a powerful graphics card in your system. To counter this problem, recently AMD has transformed their $95 AMD CPU to a 16GB GPU. It would be able to handle most of the AI workloads on Linux.
A content creator on Reddit called ttio2tech has demonstrated how we can break through the usual 2GB limit and elevate software that requires a 16GB VRAM GPU. This demo was conducted on AMD Ryzen 5 4600G which is an affordable 6-core, 12-thread, Zen 2 with 7-core Radeon Vega integrated graphics. This experiment resulted that we can easily run powerful AI tools like PyTorch and Jax on AMD’s APU. APU was capable of taking AI applications seamlessly.
The creator has generated a 512 x 512 image through AI that took almost 2 minutes to generate with default settings. He did use a DDR4 ram with that and precise details were not provided. But we can consider DDR4-3200 to get better AI performance.
It allowed many AI users to test this thing on their APU instead of buying a powerful and expensive GPU. In addition, it put a spark to AMD to create ease of access for this feature in their CPUs. For those, who want to explore the AI application world, AMD Ryzen 500 4600G could be a great option. Instead of getting a $600-$700 graphics card.
In conclusion, That clever guy on Reddit has given us a great way to turn a cheap AMD CPU into a capable GPU. Using a special patch on Linus, you would be able to run AI Tools easily with that APU. AMD Ryzen 5 4600G is already known for its cheaper price with good performance. This discovery can be a game changer for both AMD users and AMD itself to discover their chip’s power.
