4 commits

Author	SHA1	Message	Date
MaheshtheDev	457ae15259	chore(memory-graph): remove one-off cluster benchmarks ... Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>	2026-10-11 00:32:47 +00:00
Thanniru Sai Teja	ca06bfd7b6	perf(memory-graph): add engine-independent V8 benchmark, extend to 100k ... The Bun-based benchmark alone is misleading: Bun's JavaScriptCore engine optimizes Array.shift() well enough that it shows almost no difference between the old and new implementation, even though this component runs in users' browsers (predominantly V8), where the gap is real and grows with graph size. Added bench-cluster-assignments-v8.mjs, a standalone zero-dependency script runnable with plain `node`, so the V8 numbers are independently reproducible rather than just asserted. Both benchmarks now print a warning when run under the wrong engine and go up to 100k nodes.	2026-09-29 18:10:05 +05:30
Thanniru Sai Teja	fc56bb798b	perf(memory-graph): add benchmark for computeClusterAssignments ... Verifies the O(n^2) queue.shift() bottleneck in the BFS used to cluster memory-graph nodes, ahead of fixing it in a follow-up commit.	2026-09-29 17:58:02 +05:30
vorflux[bot]	851b8cfe86	Rewrite @supermemory/memory-graph with perf optimizations + consolidate consumers (#809) ... Co-authored-by: Vorflux AI <noreply@vorflux.com>	2026-03-28 19:06:27 -07:00