Performance Benchmarks
This document presents MICOS-2024 performance across different dataset scales, helping users plan resource requirements.
Test Environment
- Hardware: AMD EPYC 7742 64-Core Processor
- Memory: 256GB DDR4
- Storage: NVMe SSD
- OS: Ubuntu 22.04 LTS
- Databases: Kraken2 Standard (16GB), KneadData human_genome
Performance Data
Processing Time Comparison (hours)
Processing Time (hours)
Memory Usage (GB)
Peak Memory (GB)
Detailed Benchmark Data
| Dataset Scale | Samples | Processing Time | Memory Usage | Threads | Storage |
|---|---|---|---|---|---|
| Small | 10 | ~2 hours | 16GB | 16 | 50GB |
| Medium | 50 | ~8 hours | 32GB | 32 | 200GB |
| Large | 100 | ~15 hours | 64GB | 64 | 500GB |
| X-Large | 500 | ~72 hours | 128GB | 128 | 2TB |
Stage-by-Stage Performance
Quality Control Stage
| Step | Time % | Peak Memory | Parallelizable |
|---|---|---|---|
| FastQC | 5% | 2GB | ✓ |
| KneadData | 35% | 16GB | ✓ |
| Quality Report | 2% | 1GB | ✓ |
Taxonomic Classification Stage
| Step | Time % | Peak Memory | Parallelizable |
|---|---|---|---|
| Kraken2 | 25% | DB size | ✓ |
| Krona | 3% | 4GB | ✓ |
| BIOM Conversion | 2% | 2GB | ✓ |
Diversity Analysis Stage
| Step | Time % | Peak Memory | Parallelizable |
|---|---|---|---|
| QIIME2 | 20% | 8GB | Partial |
| Ordination | 5% | 4GB | Partial |
| Visualization | 3% | 2GB | ✓ |
Resource Planning Recommendations
Minimum Configuration
- CPU: 8 cores
- Memory: 32GB
- Storage: 100GB SSD
- Use Case: Teaching demos, small datasets
Recommended Configuration
- CPU: 32 cores
- Memory: 64GB
- Storage: 500GB NVMe SSD
- Use Case: Research projects, medium datasets
High-Performance Configuration
- CPU: 64+ cores
- Memory: 128GB+
- Storage: 2TB NVMe SSD
- Use Case: Production, large datasets
Optimization Tips
- I/O Optimization: Use NVMe SSD for FASTQ storage
- Memory Optimization: Load Kraken2 database into memory
- Parallel Optimization: Best efficiency when samples > cores
- Storage Optimization: Clean intermediate files to save space