Execution Modes

CUGA provides multiple execution modes optimized for different use cases, from fast development to precise production execution.

🎯 Mode Overview

Available Modes

Mode Selection Criteria

• Fast Mode: When speed is critical, accuracy can be slightly lower
• Accurate Mode: When precision is essential, speed can be slower
• Balanced Mode: When precision and speed are equally important
• Save & Reuse: When you have repetitive tasks that benefit from caching

🚀 Fast Mode

Purpose

Optimized for rapid development, testing, and scenarios where speed is more important than absolute precision.

Configuration

# configurations/modes/fast.toml
[features]
chat = false
task_decomposition = false
final_answer = false
thoughts = false
local_sandbox = true
code_generation = "fast"
code_output_summary = false
code_output_reflection = false
save_reuse = false
forced_apps = []

Performance Characteristics

• Execution Time: 2-5x faster than accurate mode
• Memory Usage: 30-50% less memory
• Accuracy: 75-80% of accurate mode
• Resource Usage: Optimized for throughput

🎯 Accurate Mode

Purpose

Optimized for production environments where precision, reliability, and correctness are paramount.

Configuration

# configurations/modes/accurate.toml

[features]
chat = false
task_decomposition = true
local_sandbox = false
thoughts = true
code_output_summary = true
code_generation = "accurate"
code_output_reflection = true
save_reuse = false
forced_apps = []
final_answer = true

Performance Characteristics

• Execution Time: 2-5x slower than fast mode
• Memory Usage: 30-50% more memory
• Accuracy: 98-99% precision
• Reliability: High fault tolerance

⚖️ Balanced Mode

Purpose

The best of both worlds, a middle ground between speed and precision.

Configuration

# configurations/modes/balanced.toml

[features]
chat = false
task_decomposition = true
local_sandbox = true
thoughts = false
code_output_summary = false
code_generation = "fast"
code_output_reflection = true
save_reuse = false
forced_apps = []
final_answer = true

Performance Characteristics

• Execution Time: 1-2x slower than fast mode
• Memory Usage: 20-30% more memory
• Accuracy: 90-95% precision
• Reliability: High fault tolerance

📚 Next Steps

After mastering execution modes:

1. Evaluation: Test with benchmarks
2. Development: Test and debug
3. API Integration: Extend with custom APIs
4. Save & Reuse: Explore workflow persistence