rConcentricLayout

Rust-powered, GPU-optional concentric layout engine that use for Graph Visualization Libraries like Cytoscape.js. Offload layout computation from the browser to a fast Rust backend, WebAssembly module, or GPU-accelerated WGPU pipeline.

Built for graph visualization, network topology tools, and high-performance layout workloads.

✨ Features

• ✅ Pure Rust concentric layout core
• ✅ Supports single-ring and multi-ring computations
• ✅ Designed for CPU parallelism (Rayon / custom threading)
• ✅ Optional GPU acceleration via WGPU
• ✅ Clean Rust API for layout computation
• ✅ Can be embedded in backend services or compiled to WASM

---

🚀 Purpose

rConcentricLayout allows you to:

• Execute layout computation to a Rust backend
• Run layout logic in WASM for fast, client-side graph visualization
• Use GPU acceleration when performing large or complex layout calculations
• Integrate with visualization platforms and network dashboards efficiently

---

🧠 How It Works

rConcentricLayout computes node positions using:

1. Ring levels (based on degree, user-defined, or algorithmic criteria)
2. Angular spacing
3. Radius scaling
4. Parallel CPU passes or GPU passes
5. Outputting (x, y) positions in a JSON structure compatible with Graph Visualization Libraries like Cytoscape.js

The layout can be computed in:

• Rust with multi-threaded CPU parallelism
• Rust with GPU acceleration (via WGPU compute shader)

---

📊 Benchmarking

CPU Based Parallel Computing

Number of Threads	Sample Data File	Nodes	Edges	Seconds	Milliseconds	Microseconds
2 threads	nodes_10_full_mesh.json	10 nodes	90 edges	0 s	0 ms	508 us
2 threads	nodes_100_full_mesh.json	100 nodes	9900 edges	0 s	6 ms	6008 us
2 threads	nodes_1000_random.json	1000 nodes	4000 edges	0 s	22 ms	22653 us
2 threads	nodes_2000_random.json	2000 nodes	8000 edges	0 s	89 ms	89221 us
2 threads	nodes_5000_random.json	5000 nodes	20000 edges	0 s	866 ms	866871 us
2 threads	nodes_10000_random.json	10000 nodes	40000 edges	3 s	3282 ms	3282113 us
2 threads	nodes_50000_random.json	50000 nodes	200000 edges	73 s	73045 ms	73045235 us
2 threads	nodes_100000_random.json	100000 nodes	400000 edges	295s	295919ms	295919804us
2 threads	telco_sample.json	52 nodes	220 edges	0 s	0 ms	348 us
4 threads	nodes_10_full_mesh.json	10 nodes	90 edges	0 s	0 ms	193 us
4 threads	nodes_100_full_mesh.json	100 nodes	9900 edges	0 s	5 ms	5407 us
4 threads	nodes_1000_random.json	1000 nodes	4000 edges	0 s	19 ms	19844 us
4 threads	nodes_2000_random.json	2000 nodes	8000 edges	0 s	72 ms	72126 us
4 threads	nodes_5000_random.json	5000 nodes	20000 edges	0 s	422 ms	422170 us
4 threads	nodes_10000_random.json	10000 nodes	40000 edges	1 s	1704 ms	1704142 us
4 threads	nodes_50000_random.json	50000 nodes	200000 edges	41 s	41012 ms	41012659 us
4 threads	nodes_100000_random.json	100000 nodes	400000 edges	168s	168856ms	168856157us
4 threads	telco_sample.json	52 nodes	220 edges	0 s	0 ms	408 us
8 threads	nodes_10_full_mesh.json	10 nodes	90 edges	0 s	0 ms	269 us
8 threads	nodes_100_full_mesh.json	100 nodes	9900 edges	0 s	3 ms	3203 us
8 threads	nodes_1000_random.json	1000 nodes	4000 edges	0 s	12 ms	12979 us
8 threads	nodes_2000_random.json	2000 nodes	8000 edges	0 s	42 ms	42199 us
8 threads	nodes_5000_random.json	5000 nodes	20000 edges	0 s	219 ms	219733 us
8 threads	nodes_10000_random.json	10000 nodes	40000 edges	0 s	908 ms	908255 us
8 threads	nodes_50000_random.json	50000 nodes	200000 edges	22s	22105ms	22105210us
8 threads	nodes_100000_random.json	100000 nodes	400000 edges	87s	87454ms	87454680us
8 threads	telco_sample.json	52 nodes	220 edges	0 s	0 ms	394 us
16 threads	nodes_10_full_mesh.json	10 nodes	90 edges	0 s	0 ms	398 us
16 threads	nodes_100_full_mesh.json	100 nodes	9900 edges	0 s	2 ms	2181 us
16 threads	nodes_1000_random.json	1000 nodes	4000 edges	0 s	9 ms	9709 us
16 threads	nodes_2000_random.json	2000 nodes	8000 edges	0 s	26 ms	26499 us
16 threads	nodes_5000_random.json	5000 nodes	20000 edges	0 s	118 ms	118519 us
16 threads	nodes_10000_random.json	10000 nodes	40000 edges	0 s	472 ms	472412 us
16 threads	nodes_50000_random.json	50000 nodes	200000 edges	11s	11630ms	11630157us
16 threads	nodes_100000_random.json	100000 nodes	400000 edges	45s	45904ms	45904075us
16 threads	telco_sample.json	52 nodes	220 edges	0 s	0 ms	499 us
32 threads	nodes_10_full_mesh.json	10 nodes	90 edges	0 s	0 ms	710 us
32 threads	nodes_100_full_mesh.json	100 nodes	9900 edges	0 s	4 ms	4072 us
32 threads	nodes_1000_random.json	1000 nodes	4000 edges	0 s	10 ms	10042 us
32 threads	nodes_2000_random.json	2000 nodes	8000 edges	0 s	29 ms	29730 us
32 threads	nodes_5000_random.json	5000 nodes	20000 edges	0 s	96 ms	96942 us
32 threads	nodes_10000_random.json	10000 nodes	40000 edges	0 s	397 ms	397866 us
32 threads	nodes_50000_random.json	50000 nodes	200000 edges	8 s	8436 ms	8436697 us
32 threads	nodes_100000_random.json	100000 nodes	400000 edges	34s	34646ms	34646886us
32 threads	telco_sample.json	52 nodes	220 edges	0 s	1 ms	1287 us

GPU Based Parallel Computing (A little bit of CPU for remapping the node information)

Dispatch Workgroup	WorkGroup Size	Total Threads ( Dispatch Workgroup x WorkGroup Size)	Sample Data File	Nodes	Edges	Seconds	Milliseconds	Microseconds
64	64	4096 Threads	nodes_10_full_mesh.json	10 nodes	90 edges	1 s	1262 ms	1262898 us
64	64	4096 Threads	nodes_100_full_mesh.json	100 nodes	9900 edges	1 s	1626 ms	1626835 us
64	64	4096 Threads	nodes_1000_random.json	1000 nodes	4000 edges	1 s	1100 ms	1100852 us
64	64	4096 Threads	nodes_2000_random.json	2000 nodes	8000 edges	1 s	1220 ms	1220404 us
64	64	4096 Threads	nodes_5000_random.json	5000 nodes	20000 edges	1 s	1643 ms	1643311 us
64	64	4096 Threads	nodes_10000_random.json	10000 nodes	40000 edges	2 s	2415 ms	2415705 us
64	64	4096 Threads	nodes_50000_random.json	50000 nodes	200000 edges	7 s	7937 ms	7937554 us
64	64	4096 Threads	nodes_100000_random.json	100000 nodes	400000 edges	14s	14982ms	14982001us
64	64	4096 Threads	telco_sample.json	52 nodes	220 edges	0 s	928 ms	928486 us

---

🤝 Contributions

PRs and feedback are welcome! Open an issue if you'd like to request features or propose improvements.

---

MIT