Module cache_optimizer 

Memory hierarchy and cache-aware optimization.

This module provides cache-aware optimizations for better memory performance:

• Cache modeling: Model L1/L2/L3 cache behavior
• Data layout optimization: Arrange data for cache efficiency
• Loop tiling: Optimize loop nests for cache reuse
• Prefetching: Software prefetch directives
• NUMA optimization: Optimize for non-uniform memory access

Example

use tensorlogic_infer::{CacheOptimizer, CacheConfig, TilingStrategy};

// Configure cache optimizer
let config = CacheConfig::from_system()
    .with_tiling_enabled(true)
    .with_prefetch_distance(8);

let optimizer = CacheOptimizer::new(config);

// Optimize graph for cache efficiency
let optimized = optimizer.optimize(&graph)?;

// Check cache metrics
let metrics = optimizer.estimate_cache_metrics(&optimized);
println!("Estimated cache hit rate: {:.2}%", metrics.hit_rate * 100.0);

Structs

CacheConfig

Cache configuration.

CacheMetrics

Cache metrics for a computation.

CacheOptimizer

Cache-aware optimizer.

OptimizationStats

Optimization statistics.

TilingParams

Loop tiling parameters.

Enums

AccessPattern

Memory access pattern.

CacheLevel

Cache level.

CacheOptimizerError

Cache optimization errors.

DataLayout

Data layout strategy.