RunsOn m9gd.xlarge · local NVMe vs Blacksmith 4 vCPU arm64
RunsOn does each build for 1.9–2.3× less and 1.9–2.8× faster. Both 4 vCPU arm64.
All three builds 2.0× cheaper on RunsOn: $0.0515 vs $0.1052RunsOn 2.1× faster: 9m 17s vs 19m 37s
Rust build 2.3× cheaper on RunsOn: $0.0118 vs $0.0267RunsOn 2.5× faster: 2m 04s vs 5m 04s
TypeScript build 1.9× cheaper on RunsOn: $0.0070 vs $0.0132RunsOn 2.8× faster: 50s vs 2m 21s
Docker build 2.0× cheaper on RunsOn: $0.0327 vs $0.0654RunsOn 1.9× faster: 6m 23s vs 12m 12s
RunsOn m9gd.xlarge · local NVMe
jobs ran from Amazon.com (US)
Blacksmith 4 vCPU
jobs ran from Secured Servers (US)
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 | Blacksmith | Gap |
|---|---|---|---|
| Cost | |||
| Cost per Rust build | $0.0118 | $0.0267 | 2.3× RunsOn |
| Cost per TypeScript build | $0.0070 | $0.0132 | 1.9× RunsOn |
| Cost per Docker build | $0.0327 | $0.0654 | 2.0× RunsOn |
| Price per minute | $0.0014 | $0.0050 | 3.5× RunsOn |
| Real builds, wall time | |||
| Rust build (spotify-player) | 2m 04s | 5m 04s | 2.5× RunsOn |
| TypeScript build (opencode) | 50s | 2m 21s | 2.8× RunsOn |
| Docker build (PostHog) | 6m 23s | 12m 12s | 1.9× RunsOn |
| CPU & memory | |||
| PassMark single-thread (points) | 2,463 | 1,343 | 1.8× RunsOn |
| PassMark CPU Mark, all cores (points) | 5,607 | 3,778 | 1.5× RunsOn |
| sysbench CPU, 1 thread (events/s) | 4,026 | 930 | 4.3× RunsOn |
| 7-Zip compression (MIPS) | 34,056 | 15,351 | 2.2× RunsOn |
| Memory read bandwidth | 27.9 GiB/s | 9.0 GiB/s | 3.1× RunsOn |
| Measurement | RunsOn | Blacksmith | Gap |
|---|---|---|---|
| Workspace disk (fio) | |||
| Random read, 4k QD32 (IOPS) | 87,221 | 155,975 | 1.8× Blacksmith |
| Random write, 4k QD32 (IOPS) | 43,583 | 24,132 | 1.8× RunsOn |
| Random read, 4k QD1 (IOPS) | 12,610 | 24,897 | 2.0× Blacksmith |
| QD1 read latency, p99 | 138 µs | 39 µs | 3.5× Blacksmith |
| Sequential read | 585 MiB/s | 1.7 GiB/s | 3.0× Blacksmith |
| Sequential write | 280 MiB/s | 742 MiB/s | 2.6× Blacksmith |
| Sync write, 4k + fdatasync (IOPS) | 11,264 | 7,118 | 1.6× RunsOn |
| Network, cache & Docker | |||
| Download from GitHub | 315 MiB/s | 85 MiB/s | 3.7× RunsOn |
| Cache restore, 4 GB (actions/cache) | 23s | 36s | 1.6× RunsOn |
| Cache save, 4 GB (actions/cache) | 20s | 21s | 1.1× RunsOn |
| docker pull | 5.5s | 13.7s | 2.5× RunsOn |
| Database | |||
| pgbench read/write (TPS) | 8,295 | 1,893 | 4.4× RunsOn |
| pgbench read-only (TPS) | 65,011 | 14,786 | 4.4× RunsOn |
| Queue | |||
| Queue under burst load | 21s | 15s | 1.4× Blacksmith |
- ≈ 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 themselves ran on-demand
- Cost: a managed SaaS runner against RunsOn self-hosted on your AWS account; the effort of running it was not measured