The Challenge
All benchmarks use xz -9 as the baseline — it's widely regarded as the strongest general-purpose compressor available and the standard benchmark for maximum compression.
Our previous benchmarks focused on compression ratio — how small can we make the output? But compression speed matters too. Nobody wants to wait minutes for a file to compress, especially in production pipelines.
We built a fast-mode optimizer that dramatically reduces compression time while maintaining the same compression ratios.
Results
| File | Compressed Size | vs xz -9 | Time |
|---|---|---|---|
| Linux.log | 2.2 MB | -65.4% | 0.4s |
| Apache.log | 644 KB | -74.6% | 0.5s |
| SSH.log | 71 MB | -82.3% | 4.9s |
| HDFS.log | 1,505 MB | -30.6% | 335s |
Key Takeaways
The fast-mode optimizer delivers the same compression ratios we've demonstrated in previous benchmarks — but at dramatically higher speed:
- Linux.log compresses in 0.4 seconds — 65% smaller than xz
- Apache.log compresses in 0.5 seconds — 75% smaller than xz
- SSH.log (70MB) compresses in under 5 seconds — 82% smaller than xz
- HDFS.log (1.5GB) uses the full LZMA pipeline at 335 seconds — still 31% smaller than xz
For files up to ~100MB, the optimizer path handles compression in seconds. Larger files like the 1.5GB HDFS log still use the full pipeline, but even there the results beat every traditional compressor.
What This Means
Fast mode makes Smallest.zip practical for real-time and near-real-time use cases: CI/CD artifact compression, log rotation, backup pipelines, and API-driven workflows. You no longer have to choose between compression ratio and speed.
Try It Yourself
Upload your own files at smallest.zip and see the difference. Every account starts with free credits.