High-performance Feature stores for out-of-core GNN training workloads.
- Clang 18+
- LLVM libc++
- CMake 3.16+
- Python 3.10+ (for query generation)
The project is tested on Clang with libc++. Use the provided build script to configure and compile:
./scripts/ggb_build.sh [Debug|Release|RelWithDebInfo] [--bench|--no-bench]Run the test suite using:
./scripts/ggb_run_tests.shRefer to bench/.