Performance Optimization Guide#

This guide covers performance optimization strategies for CipherSwarm v2, including hardware configuration, attack optimization, and system tuning.

Table of Contents#

Performance Optimization Guide

Overview#

CipherSwarm v2 introduces several performance enhancements:

Complexity Scoring: Intelligent attack prioritization
Keyspace Estimation: Real-time performance predictions
Enhanced Monitoring: Detailed performance metrics and analytics
Resource Optimization: Improved caching and distribution
Agent Management: Advanced hardware configuration and tuning

Hardware Optimization#

1. GPU Configuration#

GPU Selection#

Choose GPUs based on your workload:

Recommended GPUs:

High-End:
- NVIDIA RTX 4090: 24GB VRAM, excellent for large keyspaces
- NVIDIA RTX 4080: 16GB VRAM, good balance of performance/cost
- NVIDIA A100: 40/80GB VRAM, enterprise-grade performance
Mid-Range:
- NVIDIA RTX 3080: 10GB VRAM, solid performance for most tasks
- NVIDIA RTX 3070: 8GB VRAM, good for smaller keyspaces
Budget:
- NVIDIA GTX 1660 Ti: 6GB VRAM, entry-level cracking

GPU Memory Optimization#

Configure GPU memory usage:

# Agent configuration
performance:
  gpu_memory_limit: 95 # Use 95% of available GPU memory
  max_tasks: 2 # Reduce concurrent tasks for memory-intensive attacks
hashcat:
  workload: 4 # Maximum GPU utilization
  optimize: true # Enable hashcat optimizations

Temperature Management#

Maintain optimal temperatures:

# Temperature settings
hashcat:
  temp_abort: 90 # Abort at 90°C
  gpu_temp_retain: 80 # Target temperature
# Monitoring
monitoring:
  interval: 30 # Check every 30 seconds
  thermal_throttling_alert: true

2. System Configuration#

Memory Optimization#

# System memory recommendations
Minimum: 16GB RAM
Recommended: 32GB+ RAM for large campaigns
Optimal: 64GB+ RAM for enterprise deployments

# Memory settings
echo 'vm.swappiness=10' >> /etc/sysctl.conf
echo 'vm.dirty_ratio=5' >> /etc/sysctl.conf

Storage Optimization#

# Use SSD for cache and temporary files
cache_dir: /ssd/cache/CipherSwarm
temp_dir: /ssd/tmp/CipherSwarm

# Storage recommendations
Cache: 100GB+ SSD for resource caching
Temp: 50GB+ SSD for temporary files
Database: SSD for optimal query performance

Network Configuration#

# Network optimization
resources:
  download_timeout: 300
  retry_attempts: 3
  parallel_downloads: 4
  bandwidth_limit: 100MB # Optional bandwidth limiting

Attack Optimization#

1. Attack Strategy#

Complexity-Based Prioritization#

CipherSwarm v2 uses complexity scoring to optimize attack order:

Attack Complexity Levels:
  1: Very Low - Simple dictionary attacks
  2: Low - Dictionary with basic rules
  3: Medium - Mask attacks, hybrid attacks
  4: High - Complex masks, extensive rules
  5: Very High - Brute force, large keyspaces

Strategy:
  - Start with complexity 1-2 attacks
  - Progress to higher complexity based on results
  - Use previous passwords for targeted attacks

Keyspace Estimation#

Use real-time keyspace estimation for planning:

Example keyspace estimates:
Dictionary Attack: 14M words × 64 rules = 896M combinations
Mask Attack: ?u?l?l?l?l?d?d = 26 × 26^4 × 10^2 = 45.7M combinations
Brute Force: ?a^8 = 95^8 = 6.6 × 10^15 combinations

Time estimates based on agent performance:
RTX 4090: ~50 GH/s for NTLM
Estimated time: 6.6 × 10^15 ÷ 50 × 10^9 = 132,000 seconds = 36.7 hours

2. Resource Selection#

Wordlist Optimization#

Choose appropriate wordlists:

Targeted Wordlists:
  - Previous passwords (highest priority)
  - Industry-specific wordlists
  - Geographic/cultural wordlists
  - Common passwords (rockyou.txt)

Size Considerations:
  - Small wordlists (< 1M): Fast iteration, good for testing
  - Medium wordlists (1-10M): Balanced performance
  - Large wordlists (> 10M): Comprehensive but slower

Rule Selection#

Optimize rule usage:

Rule Strategies:
  Basic Rules:
    - best64.rule: 64 high-quality rules
    - OneRuleToRuleThemAll.rule: 52,000 rules (use carefully)

  Custom Rules:
    - Target-specific transformations
    - Year/date appending rules
    - Common substitutions (@ for a, 3 for e)

Performance Impact:
  - More rules = longer processing time
  - Quality over quantity approach recommended

Mask Optimization#

Design efficient masks:

Efficient Mask Patterns:
  Common Formats:
    - ?u?l?l?l?l?d?d: Capitalized word + 2 digits
    - ?l?l?l?l?d?d?d?d: Word + 4 digits
    - ?u?l?l?l?l?s: Word + special character

  Optimization Tips:
    - Use specific character classes when possible
    - Avoid overly broad masks (?a^12)
    - Consider known password policies

3. Campaign Configuration#

Attack Ordering#

Optimize attack sequence: Recommended Order:

Previous passwords (if available)
Small targeted wordlists
Dictionary + basic rules
Mask attacks (known patterns)
Dictionary + extensive rules
Hybrid attacks
Brute force (last resort)

Parallel Execution#

Configure concurrent attacks:

# Campaign settings
max_concurrent_attacks: 3 # Run multiple attacks simultaneously
attack_priority: complexity_score # Prioritize by complexity

# Agent distribution
agent_allocation: balanced # Distribute agents across attacks
min_agents_per_attack: 1 # Minimum agents per attack

Agent Performance Tuning#

1. Workload Configuration#

Workload Levels#

Configure appropriate workload:

Workload Settings (1-4):
  1: Low - 25% GPU utilization, stable but slow
  2: Medium - 50% GPU utilization, balanced
  3: High - 75% GPU utilization, good performance
  4: Maximum - 100% GPU utilization, maximum speed

Recommendations:
  - Start with workload 3
  - Increase to 4 if temperatures are stable
  - Reduce if experiencing thermal throttling

Device Selection#

Optimize device usage:

# Enable specific devices
backend_devices: 1,2 # Use GPUs 1 and 2
backend_ignore:
  cuda: false # Enable CUDA
  opencl: true # Disable OpenCL (avoid conflicts)
  hip: true # Disable HIP
  metal: true # Disable Metal

2. Performance Monitoring#

Real-time Metrics#

Monitor key performance indicators:

Key Metrics:
  - Hash rate (H/s, KH/s, MH/s, GH/s)
  - GPU utilization (target: 90%+)
  - GPU temperature (keep under 80°C)
  - Memory usage (GPU and system)
  - Task completion rate

Monitoring Tools:
  - CipherSwarm dashboard
  - Agent performance charts
  - System monitoring (nvidia-smi, htop)

Performance Baselines#

Establish performance baselines:

# Benchmark agent performance
CipherSwarm-agent benchmark

# Expected performance ranges (NTLM)
RTX 4090: 45-55 GH/s
RTX 4080: 35-45 GH/s
RTX 3080: 25-35 GH/s
RTX 3070: 20-30 GH/s

# Monitor for performance degradation
# Investigate if performance drops below 80% of baseline

3. Troubleshooting Performance Issues#

Common Performance Problems#

Low Hash Rates#

Causes:
- Thermal throttling
- Insufficient power supply
- Driver issues
- Memory limitations
Solutions:
- Improve cooling
- Upgrade power supply
- Update GPU drivers
- Reduce concurrent tasks

High Temperatures#

Causes:
- Poor ventilation
- Dust accumulation
- High ambient temperature
- Overclocking
Solutions:
- Improve case airflow
- Clean GPU fans and heatsinks
- Reduce workload setting
- Lower temperature limits

Memory Issues#

Causes:
- Large keyspaces
- Multiple concurrent attacks
- Insufficient GPU memory
Solutions:
- Reduce attack complexity
- Limit concurrent tasks
- Use agents with more memory

Resource Optimization#

1. Caching Strategy#

Agent-Side Caching#

Optimize resource caching:

# Cache configuration
resources:
  cache_dir: /ssd/cache/CipherSwarm
  max_cache: 100GB
  cleanup_interval: 3600 # Clean every hour
    # Cache policies
  wordlist_cache_ttl: 86400 # 24 hours
  rule_cache_ttl: 604800 # 7 days
  mask_cache_ttl: 604800 # 7 days

Server-Side Optimization#

# Storage service optimization
storage:
  cache_drives: [/ssd1, /ssd2]
  cache_quota: 80 # Use 80% of cache drives
# CDN integration (optional)
cdn:
  enabled: true
  provider: cloudflare
  cache_ttl: 3600

2. Resource Distribution#

Load Balancing#

Distribute resources efficiently:

# Resource distribution strategy
distribution:
  method: geographic # or "round_robin", "least_loaded"
  replication_factor: 2 # Replicate popular resources
# Bandwidth management
bandwidth:
  per_agent_limit: 100MB/s
  total_limit: 1GB/s
  priority_queue: true # Prioritize active campaigns

Prefetching#

Implement intelligent prefetching:

# Prefetch strategy
prefetch:
  enabled: true
  popular_resources: true # Cache frequently used resources
  campaign_resources: true # Prefetch campaign resources
  prediction_window: 3600 # Look ahead 1 hour

System-Level Optimization#

1. Database Performance#

Query Optimization#

-- Index optimization for common queries
CREATE INDEX idx_tasks_agent_status ON tasks(agent_id, status);
CREATE INDEX idx_campaigns_project_status ON campaigns(project_id, status);
CREATE INDEX idx_hash_items_list_status ON hash_items(hash_list_id, is_cracked);

-- Partition large tables
PARTITION TABLE hash_items BY RANGE (hash_list_id);
PARTITION TABLE tasks BY RANGE (created_at);

Connection Pooling#

# Database configuration
database:
  pool_size: 20
  max_overflow: 30
  pool_timeout: 30
  pool_recycle: 3600

2. Application Performance#

Caching Configuration#

# Application caching
cache:
  backend: cache_service # or "memory" for development
  ttl: 300 # 5 minutes default
    # Specific cache settings
  dashboard_cache_ttl: 60 # Dashboard data
  agent_status_cache_ttl: 30 # Agent status
  resource_metadata_ttl: 3600 # Resource metadata

Background Tasks#

# Background job configuration
background_jobs:
  worker_concurrency: 4
  task_routes:
    keyspace_estimation: {queue: estimation}
    resource_processing: {queue: resources}
    cleanup: {queue: maintenance}

3. Network Optimization#

Load Balancing#

# Nginx configuration
upstream cipherswarm_backend {
    server app1:8000 weight=3;
    server app2:8000 weight=2;
    server app3:8000 weight=1;
}

# Connection limits
limit_conn_zone $binary_remote_addr zone=agents:10m;
limit_conn agents 10;

SSL/TLS Optimization#

# SSL optimization
ssl_session_cache shared:SSL:10m;
ssl_session_timeout 10m;
ssl_protocols TLSv1.2 TLSv1.3;
ssl_ciphers ECDHE-RSA-AES256-GCM-SHA512:DHE-RSA-AES256-GCM-SHA512;
ssl_prefer_server_ciphers off;

Monitoring and Analytics#

1. Performance Metrics#

Key Performance Indicators#

System KPIs:

Total hash rate across all agents
Agent utilization percentage
Campaign completion rate
Resource cache hit ratio
API response times

Campaign KPIs:

Hashes cracked per hour
Time to first crack
Attack efficiency ratio
Resource utilization
Agent participation rate

Alerting#

# Performance alerts
alerts:
  low_hash_rate:
    threshold: 80% # Alert if below 80% of baseline
    duration: 300 # For 5 minutes
  high_temperature:
    threshold: 85 # Alert at 85°C
    duration: 60 # For 1 minute
  agent_offline:
    threshold: 1 # Alert immediately
    duration: 0

2. Capacity Planning#

Resource Forecasting#

# Capacity planning metrics
planning:
  agent_growth_rate: 20% # Expected monthly growth
  storage_growth_rate: 15% # Storage needs growth
  bandwidth_growth_rate: 25% # Bandwidth requirements
# Scaling thresholds
scaling:
  cpu_threshold: 80% # Scale at 80% CPU
  memory_threshold: 85% # Scale at 85% memory
  storage_threshold: 90% # Scale at 90% storage

Best Practices Summary#

1. Hardware Best Practices#

GPU Selection: Prioritize VRAM capacity for large keyspaces
Cooling: Maintain temperatures below 80°C for optimal performance
Power: Ensure adequate power supply for maximum GPU utilization
Memory: Use 32GB+ system RAM for large-scale deployments

2. Attack Best Practices#

Start Simple: Begin with low-complexity attacks
Use Intelligence: Leverage previous passwords and targeted wordlists
Monitor Progress: Use real-time metrics to adjust strategy
Optimize Order: Sequence attacks by complexity and likelihood

3. Resource Best Practices#

Cache Strategically: Cache frequently used resources locally
Size Appropriately: Balance resource size with performance needs
Validate Quality: Ensure resource quality over quantity
Monitor Usage: Track resource effectiveness and usage patterns

4. System Best Practices#

Monitor Continuously: Use comprehensive monitoring and alerting
Scale Proactively: Plan for growth and scale before bottlenecks
Optimize Regularly: Regular performance tuning and optimization
Document Changes: Track configuration changes and their impact

For additional optimization guidance:

Agent Setup Guide - Agent configuration and tuning
Attack Configuration Guide - Attack optimization strategies
Resource Management Guide - Resource optimization
Troubleshooting Guide - Performance troubleshooting