Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

iGPU with 0 VRAM with Vulkan Runtime #405

Open
thetick1234567890 opened this issue Feb 4, 2025 · 1 comment
Open

iGPU with 0 VRAM with Vulkan Runtime #405

thetick1234567890 opened this issue Feb 4, 2025 · 1 comment

Comments

@thetick1234567890
Copy link

thetick1234567890 commented Feb 4, 2025

Which version of LM Studio?
LM-Studio-0.3.9-6-x64.AppImage

Which operating system?
Linux

What is the bug?
Vulkan Runtime shows GPU but with 0 VRAM.

Can GTT memory be used with iGPU ? Update YES I see COMP engine very busy in amdgpu_top, so the GPU compute is working fine. This issue is just an issue with accurate reporting of GPU resources.

HP UEFI BIOS only allows choosing Auto, 256M, or 512M VRAM.

Various Vulkan apps work fine from https://github.com/SaschaWillems/Vulkan

Screenshots

Image

Logs

Kernel 6.12.0-12-generic #12-Ubuntu SMP PREEMPT_DYNAMIC Wed Jan 22 16:36:37 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
iGPU with AMD Ryzen 5 5600U
Device: AMD Radeon Graphics (radeonsi, renoir, LLVM 19.1.1, DRM 3.59, 6.12.0-12-generic) (0x1638)
[ 3.157034] [drm] amdgpu: 512M of VRAM memory ready.
[ 3.157040] [drm] amdgpu: 7648M of GTT memory ready.

To Reproduce
Steps to reproduce the behavior:

  1. Settings
  2. Runtime
  3. Pick Vulkan llama.ccp (Linux)
  4. Hardware -- Shows 0 VRAM for GPU
@DimitriGilbert
Copy link

Same here but it does not matter, it runs fine.
On kernel 6.10 and later you have GPP memory (dynamic vram allocation) so just remember to set the gpu offload in the model settings (usually what I forget), Bob's your uncle ;)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants