|
| 1 | +#Open Source Projects Using Metall |
| 2 | + |
| 3 | +##Collaboration Work with the ECP ExaGraph Project |
| 4 | +###miniVite |
| 5 | + |
| 6 | +miniVite is a proxy app that implements a single phase of Louvain method in distributed memory for graph community detection. |
| 7 | + |
| 8 | +miniVite has a mode that uses Metall to store a graph in persistent memory to reuse the data and reduce the overall analytics workload. |
| 9 | + |
| 10 | +For building and running miniVite with Metall, |
| 11 | +see the details located[here](https://github.com/ECP-ExaGraph/miniVite/tree/metallds2#minivite--metall-and-umap). |
| 12 | + |
| 13 | +###Ripples |
| 14 | + |
| 15 | +Ripples is a software framework to study the Influence Maximization problem developed at Pacific Northwest National Laboratory. |
| 16 | + |
| 17 | +To build Ripples with Metall, see the details located[here](./ripples.md). |
| 18 | + |
| 19 | +##HavoqGT |
| 20 | + |
| 21 | +[HavoqGT](https://github.com/LLNL/havoqgt) (Highly Asynchronous Visitor Queue Graph Toolkit) is a framework for expressing asynchronous vertex-centric graph algorithms. |
| 22 | + |
| 23 | +Same as MiniVite, HavoqGT uses Metall to store a graph in persistent memory to reuse the data and reduce the overall analytics workload. |
| 24 | + |
| 25 | +##saltatlas (DNND) |
| 26 | + |
| 27 | +[saltatlas](https://github.com/LLNL/saltatlas) is a distributed approximate k-nearest neighbors framework. |
| 28 | +saltatlas contains a distributed NNDescent algorithm implementation (DNND). |
| 29 | +DNND is designed to work with Metall to store its main data structure, which requires a heavy construction time, in persistent memory to avoid redundant data ingestion tasks. |
| 30 | + |
| 31 | +To use saltatlas DNND with Metall, see its[README](https://github.com/LLNL/saltatlas). |