RunsOn m9g.xlarge · tmpfs vs Blacksmith 4 vCPU arm64
RunsOn does the Rust and TypeScript builds for 5.5–6.0× less and 2.1–2.4× faster. Both 4 vCPU arm64.
Price per minute 3.2× cheaper on RunsOn: $0.0016 vs $0.0050the builds are not all compared on cost yet
Rust build 6.0× cheaper on RunsOn: $0.0035 vs $0.0210RunsOn 2.1× faster: 1m 52s vs 3m 59s
TypeScript build 5.5× cheaper on RunsOn: $0.0020 vs $0.0111RunsOn 2.4× faster: 48s vs 1m 55s
Docker build 10m 32s Blacksmith only · RunsOn not run
RunsOn m9g.xlarge · tmpfs
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 | $0.0035 | $0.0210 | 6.0× RunsOn |
| Cost per TypeScript build | $0.0020 | $0.0111 | 5.5× RunsOn |
| Cost per Docker build | n/a: Not run: needs more than 16 GB of RAM with tmpfs, where Docker's data root lives in RAM; even with zram swap the runner was shut down mid-build. The 32 GB tmpfs runners build it. | $0.0555 | — |
| Price per minute | $0.0016 | $0.0050 | 3.2× RunsOn |
| Real builds, wall time | |||
| Rust build (spotify-player) | 1m 52s | 3m 59s | 2.1× RunsOn |
| TypeScript build (opencode) | 48s | 1m 55s | 2.4× RunsOn |
| Docker build (PostHog) | n/a: Not run: needs more than 16 GB of RAM with tmpfs, where Docker's data root lives in RAM; even with zram swap the runner was shut down mid-build. The 32 GB tmpfs runners build it. | 10m 32s | — |
| CPU & memory | |||
| PassMark single-thread (points) | 2,470 | 1,353 | 1.8× RunsOn |
| PassMark CPU Mark, all cores (points) | 5,604 | 3,783 | 1.5× RunsOn |
| sysbench CPU, 1 thread (events/s) | 3,763 | 929 | 4.1× RunsOn |
| 7-Zip compression (MIPS) | 41,175 | 17,447 | 2.4× RunsOn |
| Memory read bandwidth | 28.2 GiB/s | 9.0 GiB/s | 3.1× RunsOn |
| Measurement | RunsOn | Blacksmith | Gap |
|---|---|---|---|
| Workspace disk (fio) | |||
| Random read, 4k QD32 (IOPS) | 2,686,147 | 253,175 | 10.6× RunsOn |
| Random write, 4k QD32 (IOPS) | 2,522,480 | 35,622 | 70.8× RunsOn |
| Random read, 4k QD1 (IOPS) | 667,201 | 32,740 | 20.4× RunsOn |
| QD1 read latency, p99 | <1 µs | 37 µs | 122.0× RunsOn |
| Sequential read | 14.7 GiB/s | 1.7 GiB/s | 8.8× RunsOn |
| Sequential write | 9.3 GiB/s | 829 MiB/s | 11.5× RunsOn |
| Sync write, 4k + fdatasync (IOPS) | 303,829 | 9,366 | 32.4× RunsOn |
| Network, cache & Docker | |||
| Download from GitHub | 166 MiB/s | 100 MiB/s | 1.7× RunsOn |
| Cache restore, 4 GB (actions/cache) | 6s | 21s | 3.5× RunsOn |
| Cache save, 4 GB (actions/cache) | 7s | 14s | 2.0× RunsOn |
| docker pull | 4.5s | 12.9s | 2.9× RunsOn |
| Database | |||
| pgbench read/write (TPS) | 9,970 | 2,948 | 3.4× RunsOn |
| pgbench read-only (TPS) | 64,583 | 24,073 | 2.7× RunsOn |
| Queue | |||
| Queue under burst load | 17s | 12s | 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)
- 2 missing on one side or both
- RunsOn at EC2 spot, its default (7-day average of the cheapest us-east-1 zone's spot price, fetched 16 Nov 2026); 9 of its jobs fell back to on-demand (AWS had no spot capacity for the type) and are priced at the spot rate over the same billed time
- Cost: a managed SaaS runner against RunsOn self-hosted on your AWS account; the effort of running it was not measured