Hello ROG Community,
I’m currently exploring the potential of using Ryzen APUs (particularly from the 5000 to 8000 series) for machine learning applications. Specifically, I am interested in configuring the VRAM allocation settings in the BIOS to support more than 16GB of VRAM. This capability would be beneficial for loading large models and datasets.
Many machine learning models can be quite large in size, and efficiently loading them often exceeds the capacity of standard CPU and RAM configurations. Being able to allocate a substantial amount of VRAM would be a more feasible way for consumers to handle these large models, potentially improving performance and reducing load times.
With the advent of newer APUs, like the ones seen in mobile PCs such as the StrixPoint APUs, and AMD’s innovative 3D V-Cache technology, it might be worth beginning to explore this capability more for the desktop range. While traditionally, this might not have made sense, it could now be highly beneficial in the context of working with machine learning models, which can be immense in size.
I have seen various discussions about increasing VRAM allocation using the “UMA Frame Buffer Size” setting in the BIOS. However, detailed user experiences about achieving allocations of 32GB, 64GB, 128GB, or even higher are sparse.
My questions are:
1. Are there any documented cases where users have successfully set the VRAM allocation to 32GB, 64GB, 128GB, or higher on Ryzen APUs?
2. Has this led to successful system boot and stable operation under such configurations?
3. Are there specific BIOS settings or configurations that need to be adjusted to achieve this?
The motherboards of particular interest for this inquiry include:
• ASUS ROG Crosshair series
• ASUS TUF Gaming series
• ASUS Prime series
The goal is to understand the feasibility and potential performance benefits of such configurations for machine learning tasks.
Any insights or experiences from the community or ASUS support team would be greatly appreciated.
Thank you!
