⚡ Bolt: Fast spatial checks for auto_k_selection and DBSCAN#144
⚡ Bolt: Fast spatial checks for auto_k_selection and DBSCAN#144teerthsharma wants to merge 1 commit into
Conversation
Replaced `distance()` (which uses `libm::sqrt`) with inline squared distance calculations (`sum < eps_sq`) in `auto_k_selection` and `DBSCAN::region_query`. This avoids the expensive square root operation in the hot inner loops of these spatial algorithms. Added NaN-safe comparisons `!(sum < eps_sq)` for robust handling. Co-authored-by: teerthsharma <78080953+teerthsharma@users.noreply.github.com>
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with New to Jules? Learn more at jules.google/docs. For security, I will only act on instructions from the user who triggered this task. |
1 participant
💡 What: Optimized spatial neighborhood queries in$O(N^2)$ checks, taking the square root for every pair is a known major performance bottleneck.
auto_k_selectionandDBSCAN::region_querywithinml::clustering.rs.🎯 Why: Previously, these algorithms calculated Euclidean distances by calling
distance(), which internally useslibm::sqrt. In spatial loops performing📊 Impact: By calculating the squared epsilon threshold (
eps_sq) upfront and breaking early inside the dimension loop if the accumulated squared sum exceeds it (!(sum < eps_sq)), we completely eliminate the square root and reduce the number of dimension iterations, yielding significantly faster clustering bounds.🔬 Measurement: Confirmed clustering logic remains correct as all topological and clustering tests (
cargo test -p aether-core) pass identically.PR created automatically by Jules for task 16304022299482913376 started by @teerthsharma