Benchmark Report
This report provides the representative benchmark snapshot currently maintained to support the whitepaper’s performance narrative. It is not a replacement for the full methodology.
Report purpose
First answer whether the result scale looks relevant to your workload, then decide whether you need to go deeper into scripts, method, or raw artifacts.
Representative results
The following results come from the maintained representative scenario: 100K reads, 150 bp, AMD Ryzen 9 5900X, Release build.
| Workload | Representative result | Meaning |
|---|---|---|
| FASTQ reading path | 1696 MB/s | Parsing and ingest throughput |
| FASTQ writing path | 1.76M reads/s | Output-stage throughput |
| Combined filtering path | 1.67M reads/s | Representative throughput under a multi-predicate QC workflow |
| Full statistics path | 302 MB/s | End-to-end statistics collection capability |
How to read it together with the whitepaper
- Return to
Performancefor the evidence framing around these numbers. - Read alongside
Architectureto judge whether the results align with the execution model. - Continue to the
Benchmark Guidewhen you need method detail.
Method reminder
These numbers are maintained representative snapshots, not guaranteed values for every environment. Storage, compression, thread count, predicate combinations, and input distribution can all change the final throughput.