quickstart →

x64arm64 at 2 vCPU4 vCPU8 vCPU

docker pullsecondslower is betterx64 · 4 vCPU

Namespace pulls an image fastest.

RunsOn is 2nd of 9, 39% above the leader. Namespace to Ubicloud: 5.2× apart.

  1. Namespace median of 2 machines 3.2s
  2. RunsOntmpfs r8a.xlarge · 32 GB 4.5s
  3. Avrea 16 GB 4.6s
  4. RunsOn median of 8 machines 5.3s
  5. Blacksmith 16 GB 7.9s
  6. AWS CodeBuild medium · 8 GB 8.1s
  7. Warpbuild 16 GB 9.7s
  8. boat 8 GB 12.6s
  9. GitHub 16 GB 16.4s
  10. Ubicloud median of 2 machines 16.9s
  11. NamespaceApple silicon arm64 only
  12. WarpbuildApple silicon arm64 only

The same image pulled on every runner, registry to disk: the download, then the layers unpacked, so the disk and the CPU count too.

Full network benchmark

The network benchmark for x64, 4 vCPU →