Build on IBM LinuxONE(s390x)
Please follow this tutorial to build and test WasmEdge on s390x hardware from source code.
Build WasmEdge
Get Source code
git clone https://github.com/WasmEdge/WasmEdge.git
Dependencies
WasmEdge requires LLVM 18 at least and you may need to install the following dependencies yourself.
As Ubuntu 22.04 and previous versions of Ubuntu do not include the llvm-18 package, you can instead use the official llvm repository to install llvm-18.
wget https://apt.llvm.org/llvm.sh
chmod +x llvm.sh
sudo ./llvm.sh 18
sudo apt-get update
sudo apt install -y software-properties-common cmake gcc g++
sudo apt install -y llvm-18-dev liblld-18-dev
Compile
Please refer to here for the descriptions of all CMake options.
cd WasmEdge
mkdir -p build && cd build
cmake -DCMAKE_BUILD_TYPE=Release .. && make -j
Run Tests
The following tests are available only when the build option WASMEDGE_BUILD_TESTS is set to ON.
Users can use these tests to verify the correctness of WasmEdge binaries.
cd <path/to/wasmedge/build_folder>
LD_LIBRARY_PATH=$(pwd)/lib/api ctest
Building Supported Plugins
Currently, the WASI-NN plugin is supported on s390x, using the GGML backend. Building the plugin with BLAS support is recommended. to build the plugin with BLAS support, first install the OpenBLAS library:
sudo apt install libopenblas-dev
Then, build WasmEdge with the following commands:
cd <path/to/wasmedge/build_folder>
make -DCMAKE_BUILD_TYPE=Release -DWASMEDGE_BUILD_PLUGINS=ON -DWASMEDGE_PLUGIN_WASI_NN_BACKEND="GGML" -DWASMEDGE_PLUGIN_WASI_NN_GGML_LLAMA_BLAS=ON ..
A guide on converting models to big endian is available at https://github.com/ggml-org/llama.cpp/blob/master/docs/build-s390x.md#getting-gguf-models