Prefer VRAM allocations on AMD GCN Vulkan GPUs#18

Open

rasbid wants to merge 1 commit intomasterfrom

codex/modify-ggml_vk_create_buffer_device-logic

Owner

rasbid commented Oct 13, 2025

Summary

prefer pure device-local Vulkan allocations on AMD GCN devices unless the host-visible fallback is explicitly enabled
log when buffer allocation falls back to host-visible memory and document the new default and override for RX 580 users

Testing

not run (not requested)

https://chatgpt.com/codex/tasks/task_e_68ebeedb9af08330903fc4ab03951052


          Prefer VRAM allocations on AMD GCN Vulkan GPUs

a509b20

rasbid added the codex label

— with

ChatGPT Codex Connector

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels