Performance

Overview

This guide covers performance optimization strategies for ComputeNet node operators, including hardware tuning, configuration optimization, and monitoring for performance issues.

Performance Metrics

Key metrics to monitor for performance:

• Job latency — Time from submission to completion
• Proof generation time — Time to generate proofs
• Attestation latency — Time to collect attestations
• Throughput — Jobs processed per unit time
• Resource utilization — CPU, memory, disk, network

Hardware Optimization

Hardware considerations for optimal performance:

• Use NVMe SSDs for low-latency storage
• Ensure adequate RAM for proof generation
• Consider GPU acceleration for proof computation
• Use high-frequency CPUs for execution
• Minimize network latency to peers

Software Configuration

Configuration options affecting performance:

• Adjust concurrent job limits based on resources
• Configure appropriate cache sizes
• Tune garbage collection parameters
• Optimize peer connection limits
• Configure log levels appropriately

Bottleneck Identification

Common performance bottlenecks and symptoms:

• CPU-bound — High CPU, slow proof generation
• Memory-bound — Swapping, OOM errors
• I/O-bound — High disk latency, queue depth
• Network-bound — High latency, packet loss

Benchmarking

Tools and techniques for benchmarking:

• Use built-in benchmarking commands
• Compare against baseline metrics
• Test with realistic workloads
• Monitor during peak usage periods