RunsOn r8a.xlarge · tmpfs vs Blacksmith 4 vCPU x64
RunsOn does each build for 4.3–5.7× less; neither is clearly faster (within 5% or inside each other's run-to-run range). Both 4 vCPU x64.
All three builds 5.1× cheaper on RunsOn: $0.0145 vs $0.0741RunsOn 1.1× faster: 7m 36s vs 8m 27s
Rust build 4.3× cheaper on RunsOn: $0.0036 vs $0.0155within run-to-run spread: 1m 52s · 1m 47s
TypeScript build 5.7× cheaper on RunsOn: $0.0021 vs $0.0121within run-to-run spread: 49s · 1m 19s
Docker build 5.3× cheaper on RunsOn: $0.0088 vs $0.0465within run-to-run spread: 4m 55s · 5m 22s
RunsOn r8a.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 | $0.0036 | $0.0155 | 4.3× RunsOn |
| Cost per TypeScript build | $0.0021 | $0.0121 | 5.7× RunsOn |
| Cost per Docker build | $0.0088 | $0.0465 | 5.3× RunsOn |
| Price per minute | $0.0016 | $0.0080 | 5.1× RunsOn |
| Real builds, wall time | |||
| Rust build (spotify-player) | 1m 52s | 1m 47s | ≈ even |
| TypeScript build (opencode) | 49s | 1m 19s | ≈ even |
| Docker build (PostHog) | 4m 55s | 5m 22s | ≈ even |
| CPU & memory | |||
| PassMark single-thread (points) | 3,816 | 4,332 | 1.1× Blacksmith |
| PassMark CPU Mark, all cores (points) | 13,502 | 14,771 | 1.1× Blacksmith |
| sysbench CPU, 1 thread (events/s) | 5,037 | 6,493 | 1.3× Blacksmith |
| 7-Zip compression (MIPS) | 38,655 | 28,802 | 1.3× RunsOn |
| Memory read bandwidth | 36.7 GiB/s | 41.7 GiB/s | 1.1× Blacksmith |
| Measurement | RunsOn | Blacksmith | Gap |
|---|---|---|---|
| Workspace disk (fio) | |||
| Random read, 4k QD32 (IOPS) | 4,201,526 | 728,035 | 5.8× RunsOn |
| Random write, 4k QD32 (IOPS) | 3,800,170 | 218,915 | 17.4× RunsOn |
| Random read, 4k QD1 (IOPS) | 1,102,306 | 71,728 | 15.4× RunsOn |
| QD1 read latency, p99 | <1 µs | 16 µs | 79.5× RunsOn |
| Sequential read | 17.8 GiB/s | 7.0 GiB/s | 2.5× RunsOn |
| Sequential write | 9.9 GiB/s | 1.7 GiB/s | 5.9× RunsOn |
| Sync write, 4k + fdatasync (IOPS) | 433,841 | 17,925 | 24.2× RunsOn |
| Network, cache & Docker | |||
| Download from GitHub | 173 MiB/s | 76 MiB/s | 2.3× RunsOn |
| Cache restore, 4 GB (actions/cache) | 5s | 11s | 2.2× RunsOn |
| Cache save, 4 GB (actions/cache) | 7s | 7s | 1.1× Blacksmith |
| docker pull | 4.5s | 7.9s | 1.8× RunsOn |
| Database | |||
| pgbench read/write (TPS) | 21,548 | 12,408 | 1.7× RunsOn |
| pgbench read-only (TPS) | 144,626 | 109,979 | 1.3× RunsOn |
| Queue | |||
| Queue under burst load | 16s | 10s | 1.6× 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 16 Nov 2026); 16 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