Install libgrape-lite with Homebrew

Project links

• Homepage
• Repository
• Upstream docs
• Package manager page

Commands and aliases

• run_app

Project history

The repository README ties libgrape-lite to GRAPE, the Graph Parallel Processing Engine described in SIGMOD and VLDB work from 2017. The project presents itself as a lightweight, open-source implementation of the GRAPE core rather than a full graph platform.

GraphScope documentation says its analytical engine originated from GRAPE, and that a lightweight version of GRAPE is open-sourced as libgrape-lite. GraphScope extends it with mutable fragments, vineyard support, and service-mode features.

Adoption history

libgrape-lite's adoption story is mostly inside Alibaba and GraphScope rather than broad end-user CLI use. The README's badge says it enables GraphScope, and GraphScope's glossary calls libgrape-lite a dependency that serves as the analytical engine core.

The v0.1.0 tag points to a 2021-10-29 commit, and subsequent tags through v0.3.5 show a small versioned release series. Homebrew packaging makes the C++/MPI graph library available outside the GraphScope source tree for users who want the lighter standalone component.

How it is used

Users build libgrape-lite with CMake, a C++11 compiler, MPI, and glog, then run example analytical applications such as single-source shortest path, weakly connected components, PageRank, local clustering coefficient, label propagation, and breadth-first search.

The project also includes LDBC benchmark support, optional GPU graph processing with CUDA and NCCL, and a GNN sampler example, so the package is most relevant to graph analytics researchers and infrastructure developers.

Why package nerds care

The package is niche but interesting because it is a standalone slice of a larger graph-computing platform. Packaging libgrape-lite separately lets users experiment with GRAPE-style graph analytics without installing all of GraphScope.

Its dependency profile is very package-manager-shaped: C++11, MPI, glog, optional jemalloc, optional CUDA/NCCL, and benchmark/examples. That makes it a small formula with heavyweight ecosystem implications.

Timeline

• 2017: GRAPE publications appeared at SIGMOD and VLDB, with the README citing Best Paper and Best Demo awards.
• 2021-10-29: The v0.1.0 tag points to a commit from this date.
• 2021: GraphScope was presented in VLDB and open sourced, with its analytical engine tied to GRAPE.
• 2026-07-01: The batch input listed Homebrew packaging for standalone libgrape-lite.

Related projects

• GraphScope is the larger graph-computing system whose analytical engine builds on GRAPE/libgrape-lite.
• Vineyard is named by GraphScope documentation as an extension point around the analytical engine.
• LDBC Graphalytics, CUDA, NCCL, MPICH, OpenMPI, glog, and GraphScope's GRAPE-JDK appear in the project's documented ecosystem.

Sources

• https://alibaba.github.io/libgrape-lite
• https://github.com/alibaba/GraphScope/blob/main/analytical_engine/README.md
• https://github.com/alibaba/GraphScope/blob/main/docs/overview/glossary.md
• https://github.com/alibaba/libgrape-lite
• https://github.com/alibaba/libgrape-lite/blob/master/README.md