Connected components, one resident pass, no self-join
You are looking at Bitcoin value-flow graph1.78M nodes · 7.37B edges in one Semurg substrate. Ask it to label connected components over 7.37B edges in one resident pass and it runs the
real query live on this box, on data you can
download and re-run yourself ▸.
Click an example in the agent rail, or ask for another seed or hop depth, e.g. WCC from 100, or label components 30-hop. Every run is measured end to end, on
this box, on the real dataset. Nothing is pre-computed.
the live Bitcoin value-flow graph, transactions and addresses linked by value-flows (prev-tx to tx to address), parsed straight off the P2P network · Bitcoin mainnet P2P (our own block-parser node, no bitcoind, no REST). Weakly-connected-components by label propagation, run as a single co-located image sweep instead of a self-join per round. Ask the agent (right); the number is measured here, now, on data you can download.
the relational leader
A self-join every round. Connected components by recursive query is a self-join re-materialized once per propagation round, paying a full index join for every round until the labels settle.
semurg, one engine
One resident pass. The whole graph is labelled in a single co-located image sweep with no join and no re-read, because the compute runs on the data.
Rounds times index-join collapses to one sweep.
Bitcoin value-flow graph: 1.78M nodes, 7.37B edges. Measured live on this box. Download it and re-run it yourself.
ask the agent to label components, the live throughput renders here
This is a RAW win, and we say so plainly: connected components on a relational store is a self-join repeated once per propagation round, paying an index join every round. Here it is a single resident pass over the in-memory adjacency, because the compute runs on the data. Every number is measured live on the AVX-512 box, on a public dataset you can download and re-run yourself, the same engine that also serves graph traversal, full-text and vectors.