Benchmark Overview

FastQTools keeps a compact benchmark summary so users can estimate how the project fits common FASTQ QC workloads without reading raw CI artifacts first.

Representative results

These results are the current maintained snapshot for 100K reads (150 bp) on an AMD Ryzen 9 5900X using a Release build.

Workload	Representative result	What it represents
FASTQ read path	1696 MB/s	Parsing and ingest throughput
FASTQ write path	1.76M reads/s	Output-side throughput in the maintained benchmark set
Combined filtering pass	1.67M reads/s	A realistic QC pass with multiple predicates enabled
Full statistics pass	302 MB/s	End-to-end metrics collection

How to read these numbers

• They are representative, not guaranteed minima or maxima.
• Actual throughput depends on storage, compression level, thread count, filter mix, and read length distribution.
• The table is intentionally small: it is meant to answer “is this in the right performance range for my workload?” before you dive into deeper reports.

Benchmark workload

• Dataset shape: synthetic FASTQ data, 100K reads, 150 bp per read
• Environment: Linux on AMD Ryzen 9 5900X
• Build profile: Release
• Focus: routine FASTQ QC tasks, not cross-tool marketing comparisons

Benchmark artifacts and tooling

• This page is the stable, curated overview for public readers.
• CI-generated benchmark artifacts may be published under docs/benchmark-reports/ when available.
• Benchmark tooling lives in scripts/tools/performance/ inside the repository.

When benchmarks matter most

Benchmarks are most useful when you are deciding whether FastQTools is a good fit for:

1. pre-alignment filtering in existing QC pipelines,
2. repeated batch statistics over many FASTQ files, or
3. embedding FASTQ processing into a larger C++ application.

If you need command syntax next, go to the CLI Reference. If you need integration details, go to the API Overview.