Numbers
Every figure on this site is read from bench/site-numbers.json. Each row names the command that produced it.
Linux
bench/linux-three-way.sh, p16s, Core Ultra 7 165H, Linux 6.17, Docker 29.1.3, Podman 4.9.3 rootless, 2026-11-07. Source: docs/PERFORMANCE.md.
| Image | Metric | corail | Docker | Podman |
|---|---|---|---|---|
| alpine | warm run --rm /bin/true, median of 20 | 6 ms | 214 ms | 110 ms |
| alpine | exec /bin/true, median of 20 | 2 ms | 53 ms | 105 ms |
| alpine | host footprint of 50 idle sleeps | 54 MB | 1.2 GB | 107 MB |
| busybox | warm run --rm /bin/true, median of 20 | 7 ms | 289 ms | 147 ms |
| busybox | exec /bin/true, median of 20 | 4 ms | 69 ms | 119 ms |
| busybox | host footprint of 50 idle sleeps | 53 MB | 1.2 GB | 106 MB |
| debian | warm run --rm /bin/true, median of 20 | 6 ms | 265 ms | 135 ms |
| debian | exec /bin/true, median of 20 | 2 ms | 61 ms | 116 ms |
| debian | host footprint of 50 idle sleeps | 53 MB | 1.2 GB | 106 MB |
macOS
bench/mac-image-start.py, bench/mac-image-rest.py, bench/mac-image-idle.py, Apple M4 Pro, macOS 26.6.2, Docker Desktop 4.98.0 (engine 29.9.2), Podman 6.1.4 applehv, 4 vCPU, 2 GiB, 2026-11-16. Source: docs/PERFORMANCE.md, section macOS three-way, refreshed 2026-11-16.
| Image | Metric | corail | Docker | Podman |
|---|---|---|---|---|
| alpine | warm run --rm /bin/true, median of 20 | 5.4 ms | 132 ms | 152 ms |
| alpine | exec /bin/true, median of 20 | 5.2 ms | 33 ms | 115 ms |
| alpine | host footprint of 50 idle sleeps | 196 MB | 1.5 GB | 1.3 GB |
| busybox | warm run --rm /bin/true, median of 20 | 5.2 ms | 140 ms | 155 ms |
| busybox | exec /bin/true, median of 20 | 4.2 ms | 31 ms | 114 ms |
| busybox | host footprint of 50 idle sleeps | 190 MB | 1.4 GB | 1.3 GB |
| debian | warm run --rm /bin/true, median of 20 | 6.6 ms | 145 ms | 159 ms |
| debian | exec /bin/true, median of 20 | 4.3 ms | 31 ms | 113 ms |
| debian | host footprint of 50 idle sleeps | 340 MB | 1.4 GB | 1.3 GB |
Targets and the control plane
A target is a target, not a result. The measured column is from just bench-check with SITE_NUMBERS=1; a row says "not re-recorded" when that run has no figure for it.
| Metric | Measured | Target | Command, commit |
|---|---|---|---|
| warm run --rm /bin/true, median of 10 | 5 ms | ≤ 9 ms | just bench-check (run --rm alpine -- /bin/true, median of 10), p16s d21ab26 |
| exec /bin/true, median of 10 | 2 ms | ≤ 15 ms (guard) | just bench-check (exec /bin/true, median of 10), p16s d21ab26 |
| recify apply to Running | 6 ms | ≤ 50 ms, guarded | bench/coraild.sh, p16s d21ab26 |
| kill to recovered | 8 ms | ≤ 300 ms, guarded | bench/coraild.sh, p16s d21ab26 |
| coraild RSS at 50 atolls | 3.6 MB | < 16 MB | bench/coraild.sh, p16s d21ab26 |
The comparison against Kubernetes (bench/vs-k8s.sh) has not been run for this site, so no Kubernetes figure appears here.
How we measure
- Tools are sampled interleaved, and the tool that goes first rotates each round.
- Three warm-up rounds are discarded. Latency rows use at least 20 samples and report the median.
- The image is pulled on every side before timing starts.
- A verdict needs a paired sign test, |z| ≥ 1.96, with at least 80% agreement (
bench/ab-verdict.sh). - Memory is resident size summed over the daemon and each container's supervisor.
- A network-bound figure, such as a cold pull, is only an absolute hang guard. It measures the link, not the code.
- Rows where corail is worst are flagged as findings, not dropped.
Rules in full: bench/results/latest-p16s.md and docs/PERFORMANCE.md.