Expand description
Advanced-advanced memory optimization system for ODE solvers
This module provides cutting-edge memory management optimizations including:
- Predictive memory allocation based on problem characteristics
- Multi-level memory hierarchy optimization (L1/L2/L3 cache, RAM, GPU)
- Adaptive memory layout reorganization for maximum cache efficiency
- Real-time memory usage monitoring and optimization
- Zero-copy buffer management and memory-mapped operations
- NUMA-aware memory allocation for multi-socket systems
Structs§
- Advanced
Memory Optimizer - Advanced-advanced memory optimization manager
- Allocation
Event - Memory allocation event for learning
- Allocation
Pattern - Allocation pattern information
- Allocation
Predictor - Predictive memory allocation engine
- Allocation
Strategy - Buffer
Reuse Statistics - Buffer reuse statistics
- Buffer
Usage Stats - Buffer usage statistics
- Cache
Aware Algorithm Selector - Cache-aware algorithm selector
- Cache
Blocking Manager - Cache blocking strategy manager
- Cache
Hierarchy Info - Cache hierarchy information
- Cache
Optimizer - Cache optimization system
- Cache
Performance Metrics - Cache performance metrics
- CpuAffinity
Manager - CPU affinity manager
- Data
Layout Optimizer - Data layout optimizer
- Dimension
Analyzer - Dimension analyzer for problem size characteristics
- Fragmentation
Analyzer - Fragmentation analyzer
- GpuBuffer
- GPU memory buffer for heterogeneous computing
- L1Cache
Buffer - L1 cache-optimized buffer (typically 32KB per core)
- L2Cache
Buffer - L2 cache-optimized buffer (typically 256KB-1MB per core)
- L3Cache
Buffer - L3 cache-optimized buffer (typically 8-32MB shared)
- Memory
Hierarchy Manager - Multi-level memory hierarchy management
- Memory
Leak Detector - Memory leak detector
- Memory
Placement Policy - Memory placement policy
- Memory
Requirements - Memory
Usage Statistics - Memory usage statistics
- Memory
Usage Tracker - Memory usage tracker
- Mmap
Buffer - Memory-mapped buffer for large datasets
- Numa
Topology - NUMA topology information
- Numa
Topology Manager - NUMA topology management
- Optimization
Plan - Optimized
Memory Region - Performance
Counters - Performance counters
- Performance
Impact - Performance impact measurement
- Prediction
Accuracy Tracker - Prediction accuracy tracker
- Prefetch
Pattern Optimizer - Prefetch pattern optimizer
- Problem
Characteristic Analyzer - Problem characteristic analysis for prediction
- Problem
Characteristics - RamBuffer
- Main memory buffer with NUMA optimization
- Real
Time Memory Monitor - Real-time memory monitoring system
- Sparsity
Analyzer - Sparsity pattern analyzer
- Stiffness
Analyzer - Stiffness characteristic analyzer
- Temporal
Analyzer - Temporal access pattern analyzer
- Zero
Copy Buffer - Zero-copy buffer implementation
- Zero
Copy Buffer Pool - Zero-copy buffer pool for efficient data transfer
Enums§
- Access
Mode - Memory access modes for mmap
- Access
Pattern - Memory access patterns for optimization
- Buffer
Reuse Strategy - Cache
Strategy - Coherency
State - Memory coherency states
- GpuMemory
Type - GPU memory types
- Memory
Layout - Memory layout optimization strategies
- Memory
Tier - Memory
Type - Memory type classification
- Numa
Placement - Prefetch
Schedule - Prefetch
Strategy - Memory prefetch strategies
- Replacement
Policy - Cache replacement policies
- Sharing
Strategy - Cache sharing strategies for L3
- Temporal
Access Pattern