Comparison
Graviton vs Intel for PostgreSQL
AWS Graviton (Arm) vs Intel (x86) EC2 instances running self-managed PostgreSQL — head-to-head throughput, latency, and cost-efficiency on a real OLTP benchmark.
Verdict
For this PostgreSQL workload, m7g.2xlarge (arm64) is the best value at 217,449 req/s per $/hr — a 1% cost-efficiency lead over c8i.2xlarge. The highest raw throughput comes from c8i.2xlarge at 85,752 req/s. Across architectures, the best Arm option (m7g.2xlarge) leads the best x86 option (c8i.2xlarge) by 1% on cost-efficiency.
Side by side
| Instance | Arch | $/hr | RPS | p95 ms | RPS/$ | RPS/vCPU |
|---|---|---|---|---|---|---|
| m7g.2xlarge | arm64 | $0.3483/hr | 75,741 | 1.82 | 217,449 | 9,468 |
| c8i.2xlarge | x86-64 | $0.3968/hr | 85,752 | 1.73 | 216,132 | 10,719 |
| m8g.2xlarge | arm64 | $0.3810/hr | 81,055 | 1.75 | 212,767 | 10,132 |
| m7i.2xlarge | x86-64 | $0.4251/hr | 71,930 | 1.81 | 169,201 | 8,991 |
Measured at a 10 GB dataset on gp3-baseline storage. Open the sizing tool to change the dataset size, disk, or metric, or browse the full benchmark data.
Frequently asked questions
Which is better for PostgreSQL, m8g.2xlarge or c8i.2xlarge or m7g.2xlarge or m7i.2xlarge?
On cost-efficiency, m7g.2xlarge leads at 217,449 req/s per $/hr. For raw throughput, c8i.2xlarge is fastest at 85,752 req/s. Across architectures, the best Arm option (m7g.2xlarge) leads the best x86 option (c8i.2xlarge) by 1% on cost-efficiency. Choose by whether you optimize for throughput per dollar or peak capacity.
m7g.2xlarge vs c8i.2xlarge: what's the difference for PostgreSQL?
m7g.2xlarge delivers 75,741 req/s at 217,449 req/s per $/hr; c8i.2xlarge delivers 85,752 req/s at 216,132 req/s per $/hr. That is a 1% cost-efficiency edge for m7g.2xlarge on this workload.
What workload and configuration were these numbers measured on?
A mixed OLTP pattern — 90% primary-key reads, 10% single-row inserts — at 32 concurrent connections on PostgreSQL 17 in us-east-1, with a 10 GB dataset on a gp3-baseline gp3 volume. The benchmark harness is open source.