Latest (0.1.27) docker image with ROCm works for me on Ryzen 5600G with 8GB VRAM allocation. Prompt processing is 2x faster than with CPU. Generation runs at max speed even if CPU is busy running other processes. I am on Fedora 39. Container setup: HSA_OVERRIDE_GFX_VERSION=9.0.0

Use this to install ROCm on Tumbleweed. Does NOT talk about Pytorch, the tag is for completeness.

SHOWS EXACTS SPECS OF CPU OR APU INCLUDING GRAPHICS

Compare Distinct graphics cards with AMD APUs. Result: Buy Ryzen 5700G, or AMD Ryzen 5 5600GT as GPU is the same. BUT Ryzen 5700G has 25% better CPU

This guy claims he got both gfx900 and gfx902 working. I asked to try my test.

Database of AMD GPUs. Here we can see that 2500U is Vega 8 mobile , which is GCN 5.0. This is only supported in ROCm up to 4.5.2!!

Rusticl seems equivalent of rocm.

This site seems to claim there is a known fix. Looks like someone is fixing it???

Installing PyTorch for ROCm - this document claims gfx900 compatibility

Looks like gfx900, gfx902, gfx909 and gfx90c are the same

Suggested git-build of pytorch on gfx90c FAILED for me

LM Studio can be installed on Linux with APU or GPU (looks like it needs the AI CPU though??) and run LLM. Install on Laptop and test if it works.

If you have AMD GPU as I do then you can grab PCI ID for the device with lspci command executed with -D flag (shows PCI doamin) and read the following file cat /sys/bus/pci/devices/${pci_slot}/mem_info_vram_total, it contains GPU VRAM size in bytes.

Top of the guide describing ROCm on Linux. There are 2 core approaches: Using RPM (Package manager), or using AMD installer. I should use Package manager. Also single-version vs. multi-version. I should use single-version, latest.

The links in “How to guide“ provide instructions that are hopeful. Maybe start with those instructions!

This guy seems to claim ROCM can run on Tumbleweed using Distrobox. But what is distrobox?

This guy claims successful installation of ROCm on Ubuntu - this seems to be workable for Tumbleweed as well. See the comment “nav9 commented on Jul 16, 2023“

Train LLM on AMD APU. In this scenario, we’ll use an APU because most laptops with a Ryzen CPU include an iGPU; specifically, this post should work with iGPUs based on the “GCN 5.0” architecture, or “Vega” for friends. We’ll use an AMD Ryzen 2200G in this post, an entry-level processor equipped with 4C/4T and an integrated GPU.

UMA buffer size is the size of memory used by APU. It is set on the motherboard, often limited to 2GB. But LLM AI could use 16GB or more.