RunsOn i7i.xlarge · local NVMe vs GitHub 4-core x64
RunsOn does each build for 2.9–6.6× less; on time, RunsOn is faster on the Docker build and tied on the Rust and TypeScript builds. Both 4 vCPU x64.
All three builds 3.4× cheaper on RunsOn: $0.0735 vs $0.2498RunsOn 1.4× faster: 12m 57s vs 18m 27s
Rust build 6.1× cheaper on RunsOn: $0.0077 vs $0.0466about the same time: 3m 06s · 3m 47s
TypeScript build 6.6× cheaper on RunsOn: $0.0041 vs $0.0268about the same time: 1m 17s · 1m 56s
Docker build 2.9× cheaper on RunsOn: $0.0618 vs $0.1764RunsOn 1.5× faster: 8m 34s vs 12m 44s
RunsOn i7i.xlarge · local NVMe
jobs ran from Amazon.com (US)
GitHub 4-core
jobs ran from Microsoft (US, CA)
Side by side
Every measurement both runners share, cost first. The dot marks the better value; gap is the larger over the smaller.
| Measurement | RunsOn | GitHub | Gap |
|---|---|---|---|
| Cost | |||
| Cost per Rust build | $0.0077 | $0.0466 | 6.1× RunsOn |
| Cost per TypeScript build | $0.0041 | $0.0268 | 6.6× RunsOn |
| Cost per Docker build | $0.0618 | $0.1764 | 2.9× RunsOn |
| Price per minute | $0.0019 | $0.0120 | 6.3× RunsOn |
| Real builds, wall time | |||
| Rust build (spotify-player) | 3m 06s | 3m 47s | ≈ even |
| TypeScript build (opencode) | 1m 17s | 1m 56s | ≈ even |
| Docker build (PostHog) | 8m 34s | 12m 44s | 1.5× RunsOn |
| CPU & memory | |||
| PassMark single-thread (points) | 2,848 | 2,370 | 1.2× RunsOn |
| PassMark CPU Mark, all cores (points) | 7,158 | 6,130 | 1.2× RunsOn |
| sysbench CPU, 1 thread (events/s) | 2,866 | 3,670 | 1.3× GitHub |
| 7-Zip compression (MIPS) | 19,709 | 16,137 | 1.2× RunsOn |
| Memory read bandwidth | 19.8 GiB/s | 20.9 GiB/s | 1.1× GitHub |
| Measurement | RunsOn | GitHub | Gap |
|---|---|---|---|
| Workspace disk (fio) | |||
| Random read, 4k QD32 (IOPS) | 150,053 | 19,587 | 7.7× RunsOn |
| Random write, 4k QD32 (IOPS) | 82,488 | 11,712 | 7.0× RunsOn |
| Random read, 4k QD1 (IOPS) | 12,069 | 6,843 | 1.8× RunsOn |
| QD1 read latency, p99 | 307 µs | 251 µs | 1.2× GitHub |
| Sequential read | 1,010 MiB/s | 394 MiB/s | 2.6× RunsOn |
| Sequential write | 788 MiB/s | 394 MiB/s | 2.0× RunsOn |
| Sync write, 4k + fdatasync (IOPS) | 12,095 | 3,262 | 3.7× RunsOn |
| Network, cache & Docker | |||
| Download from GitHub | 287 MiB/s | 91 MiB/s | 3.1× RunsOn |
| Cache restore, 4 GB (actions/cache) | 13s | 41s | 3.1× RunsOn |
| Cache save, 4 GB (actions/cache) | 9s | 30s | 3.3× RunsOn |
| docker pull | 7.1s | 16.4s | 2.3× RunsOn |
| Database | |||
| pgbench read/write (TPS) | 6,645 | 2,797 | 2.4× RunsOn |
| pgbench read-only (TPS) | 54,071 | 20,013 | 2.7× RunsOn |
| Queue | |||
| Queue under burst load | 19s | 4s | 4.8× GitHub |
- ≈ even: under 5% apart, or a build inside both runners' run-to-run spread
- Queue: identical jobs queued at the same instant (burst suite)
- RunsOn at EC2 spot, its default (7-day average of the cheapest us-east-1 zone's spot price, fetched 5 Oct 2026); the jobs ran on spot
- Cost: a managed SaaS runner against RunsOn self-hosted on your AWS account; the effort of running it was not measured