Skip to content

docs: add Windows HIP build instructions and troubleshooting #285

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.

Sign up for GitHub

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

Merged
hydai merged 1 commit into from
Jan 28, 2026
+24 −0
Merged

docs: add Windows HIP build instructions and troubleshooting #285

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 24 additions & 0 deletions docs/contribute/source/plugin/wasi_nn.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -328,6 +328,14 @@ Then developers can build by following the steps.
& "<the ninja-build folder>\ninja.exe" -C build
```

- If you want to enable HIP (AMD GPU):

```console
# HIP ENABLE:
& "C:\Program files\CMake\bin\cmake.exe" -Bbuild -GNinja -DCMAKE_BUILD_TYPE=Release -DWASMEDGE_PLUGIN_WASI_NN_BACKEND=ggml -DWASMEDGE_PLUGIN_WASI_NN_GGML_LLAMA_HIP=ON -DWASMEDGE_USE_LLVM=OFF .
& "<the ninja-build folder>\ninja.exe" -C build
```

#### Execute the WASI-NN plugin with the llama example on Windows

1. Set the environment variables
Expand All @@ -345,6 +353,22 @@ Then developers can build by following the steps.
wasmedge --dir .:. --env llama3=true --env n_gpu_layers=100 --nn-preload default:GGML:AUTO:Meta-Llama-3-8B-Instruct.Q5_K_M.gguf wasmedge-ggml-llama.wasm default
```

#### Troubleshooting: AMD Radeon Integrated Graphics (Windows)

If you are building the GGML backend on Windows with an integrated AMD GPU (e.g., Radeon 780M / gfx1103) and encounter `rocBLAS` or initialization errors, you may need the following workarounds:

1. **Set the Architecture Override:**
The RDNA 3 integrated graphics require an override to match the available ROCm kernels.
```console
$env:HSA_OVERRIDE_GFX_VERSION = "11.0.0"
```

2. **Manual Library Linking (ROCm 6.x):** The ROCm installer on Windows may not create the necessary symlinks for `gfx1103`. If you see errors related to missing `TensileLibrary` files:

* Navigate to your ROCm installation folder (e.g., `C:\Program Files\AMD\ROCm\6.x\bin\rocblas\library`).
* Copy the `gfx1100` files and rename them to `gfx1103`.
* Example: Copy `TensileLibrary_lazy_gfx1100.dat` -> `TensileLibrary_lazy_gfx1103.dat`

### Appendix for llama.cpp backend

We also provided the pre-built ggml plugins on the following platforms:
Expand Down
Loading