Struct SimdTripleMatcher 

pub struct SimdTripleMatcher { /* private fields */ }

SIMD-optimized triple matcher with full SciRS2-core integration

Uses CPU vector instructions and SciRS2-core’s advanced array operations to match multiple triples against patterns simultaneously. This provides significant speedup for large-scale pattern matching operations in SPARQL query evaluation.

Features:

• Array-based SIMD operations with SciRS2-core
• Parallel processing for large batches (with “parallel” feature)
• Built-in performance metrics tracking
• Platform-adaptive (AVX2/AVX-512/NEON via scirs2-core)

Implementations

impl SimdTripleMatcher

pub fn new() -> Self

Create a new SIMD triple matcher with default settings and metrics

pub fn with_chunk_size(chunk_size: usize) -> Self

Create a matcher with custom chunk size

pub fn stats(&self) -> MatcherStats

Get performance statistics for this matcher

pub fn reset_stats(&self)

Reset all performance statistics

Note: Resetting metrics is not supported by scirs2-core’s Counter and Timer, so we only reset the atomic counters. Create a new matcher for fresh stats.

pub fn match_batch( &self, pattern: &TriplePattern, triples: &[Triple], ) -> Result<Vec<usize>>

Match a batch of triples against a pattern using SIMD

This is the primary entry point for SIMD-optimized pattern matching. It processes triples in chunks using SIMD operations for maximum throughput.

Arguments

• pattern - The triple pattern to match against
• triples - The triples to check for matches

Returns

A vector of indices indicating which triples matched the pattern

pub fn estimate_selectivity( &self, pattern: &TriplePattern, _total_triples: usize, ) -> f32

Estimate selectivity of a pattern for query optimization

Returns a value between 0.0 (no matches) and 1.0 (all match)

Trait Implementations

impl Default for SimdTripleMatcher

fn default() -> Self

Returns the “default value” for a type.