- FusedMM: FusedMM is a general-purpose kernel that unifies SDDMM and SpMM kernels for efficient computations of Graph Embedding and Graph Neural Networks. This software is written in C/C++ using OpenMP support. It also supports SIMD vectorization for Intel, AMD, and ARM architectures.
- Force2Vec: This is an unsupervised graph embedding method that can effectively utilize the computing resources and generate good quality embeddings. This software is written in C++ and OpenMP. This also employs AVX512 instructions for SIMD vectorization.
- BatchLayout: BatchLayout is a parallel algorithm for large-scale graph visualizations. It employs a spring-electrical model to generate the high quality layouts of graphs. This software is written using C++ and OpenMP.
- HipMCL: HipMCL is a high-performance parallel algorithm for large-scale network clustering. HipMCL parallelizes popular Markov Clustering (MCL) algorithm. Written using C++, MPI and OpenMP. Scales up to 130,000 cores on a Cray XC40 supercomputer.
- Combinatorial BLAS: A distributed-memory parallel graph library offering a small but powerful set of linear algebra primitives specifically targeting graph analytics. Written using C++, MPI and OpenMP. Includes implementations of distributed-memory cardinality and weighted matching and distributed-memory Reverse Cuthill McKee (RCM) ordering.
- MS-BFS-Graft: A shared-memory parallel implementation of bipartite cardinality matching algorithms. Written using C++ and OpenMP.
- FlowMatch: A biconductor package for matching cell clusters and building templates from flow cytometry data. Written in R and C++. More than 7500 downloads. [bioconductor link]
- FlowVS: A biconductor package to stabilize variance in flow cytometry and microarray data. Written in R and C++. More than 6500 downloads. [bioconductor link]