Skip to content

BUILDING.md

Latest commit

 

History

History
95 lines (73 loc) · 2.52 KB

BUILDING.md

File metadata and controls

95 lines (73 loc) · 2.52 KB
Raw

Building FasterForest

Prerequisites

  • JDK 22+ (required for Panama FFM native access)
  • Gradle (wrapper included, no install needed)

For native library builds:

  • CMake 3.16+
  • GCC (Linux) or MSVC (Windows)

Quick Build (Java only)

If you don't need to recompile the native C library (pre-compiled binaries are already included in src/main/resources/native/):

./gradlew clean assemble

The JAR is output to build/libs/FasterForest-<version>.jar.

Building with Native Library

The native library provides C-based forest prediction via Panama FFM. Pre-compiled binaries for Linux and Windows (x86_64) are committed to the repo, so you only need to rebuild them if you change the C source code.

Linux x86_64

./build-linux.sh

This script:

  1. Builds the native libfasterforest.so via CMake
  2. Copies it to src/main/resources/native/linux-x86_64/fasterforest.so
  3. Builds the JAR with ./gradlew clean assemble

Or manually:

cd native
mkdir -p build && cd build
cmake ..
make
cp libfasterforest.so ../../src/main/resources/native/linux-x86_64/fasterforest.so
cd ../..
./gradlew clean assemble

Windows x86_64

With MSVC (Visual Studio):

cd native
mkdir build && cd build
cmake .. -G "Visual Studio 17 2022"
cmake --build . --config Release
copy Release\fasterforest.dll ..\..\src\main\resources\native\windows-x86_64\fasterforest.dll
cd ..\..
gradlew clean assemble

With MinGW:

cd native
mkdir build && cd build
cmake .. -G "MinGW Makefiles"
cmake --build .
copy libfasterforest.dll ..\..\src\main\resources\native\windows-x86_64\fasterforest.dll
cd ..\..
gradlew clean assemble

Cross-platform JAR

To build a JAR containing native libraries for both platforms:

  1. Build the native library on each platform (or cross-compile)
  2. Place the binaries in the resource tree:
    • src/main/resources/native/linux-x86_64/fasterforest.so
    • src/main/resources/native/windows-x86_64/fasterforest.dll
  3. Run ./gradlew clean assemble

The JAR bundles all native libraries found in the resource tree. At runtime, NativeLoader extracts the correct one for the current OS/arch.

Notes

  • AVX2 SIMD code is compiled only on x86_64. Runtime CPUID detection falls back to scalar if AVX2 is not available on the target machine.
  • The native library is optional. If loading fails, FasterForest falls back to pure Java prediction automatically.
  • For development, you can point to a local native library with: -Dfasterforest.native.path=/path/to/libfasterforest.so