You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
EGGPU is a GPU-accelerated network analysis library that supports essential functions such as betweenness centrality, k-core centrality, and single-source shortest path. Built on top of the **EasyGraph** library, EGGPU delivers a user-friendly Python API while achieving remarkable speedups for large-scale network analysis.
8
+
9
+
EGGPU is engineered with a three-layer architecture:
10
+
- User Interface Layer: Developed in Python, this layer offers intuitive and easy-to-use APIs for end users.
11
+
- Middleware Layer: Constructed in C++, this layer shares memory space with the Computation Layer and serves as the binding agent. It also provides a graph container responsible for graph loading, storage, and format conversion.
12
+
- Computation Layer: Implemented in CUDA C/C++, this layer primarily executes the GPU-based network analysis functions, including betweenness centrality, k-core centrality, and SSSP.
0 commit comments