m8a.large

5th Generation AMD EPYC processors (codename Turin) with a maximum frequency of 4.5 GHz Built on the AWS Nitro System, offers up to 30% higher performance and 45% more memory bandwidth than M7a instances, and supports up to 75 Gbps networking bandwidth and Instance Bandwidth Configuration (IBC)

available for self-hosted runners on RunsOn

cpu speed · passmark single-thread

3,766pts

AMD EPYC 9R45 (x86_64)
averaged over many days · higher is faster

vcpus

physical threads

memory

7.6 GiB

ram

clock speed

Up to 4.5 GHz

max frequency

architecture

x86_64

EC2 instance family: m8a

What the family is built for, the silicon underneath, and how to ask RunsOn for it.

m8a general purpose

cpu type

AMD EPYC 9R45 Processor — 5th Generation AMD EPYC processors (codename Turin) with a maximum frequency of 4.5 GHz

architecture

x86_64 · x86_64

highlights

Built on the AWS Nitro System, offers up to 30% higher performance and 45% more memory bandwidth than M7a instances, and supports up to 75 Gbps networking bandwidth and Instance Bandwidth Configuration (IBC)

released

Generally available on October 8, 2025

.github/workflows/ci.yml yaml

jobs:
  build:
    runs-on:
      - runs-on=${{ github.run_id }}
      - family=m8a
      - cpu=2

the label is a query — each line one constraint, resolved at launch. job labels →

Availability & price by region

Spot is the average price RunsOn pays in your own AWS account — no per-minute markup. The bar shows how far under on-demand each region runs; the badge is the observed spot-interruption frequency.

N. Virginia

us-east-1

spot · avg $0.0531/hr

on-demand$0.1217

savings vs on-demand63%

spot interrupt freq >20%

Oregon

us-west-2

best spot

spot · avg $0.0374/hr

on-demand$0.1217

savings vs on-demand68%

spot interrupt freq 10-15%

Ireland

eu-west-1

spot · avg $0.0648/hr

on-demand$0.1357

savings vs on-demand—

spot interrupt freq —

Frankfurt

eu-central-1

spot · avg $0.0849/hr

on-demand$0.1458

savings vs on-demand55%

spot interrupt freq <5%

Spot prices are trailing averages per region and move with the market. On-demand is the list reference. Interrupt frequency is AWS's published spot-advisor band for this instance type.

About these benchmarks

CPU is the Passmark single-thread metric, run on a live m8a.large instance and averaged over many days so a single noisy run can't skew the score.

Pricing is normalized to the hourly rate AWS bills, then shown per region. Spot is the trailing average; savings is measured against the same region's on-demand list price.

This instance is launched the exact same way any RunsOn job launches it — ephemeral EC2 in your account, one CloudFormation stack, nothing to babysit.

metricpassmark · single-thread

samples21 runs

regions4 · us-east-1, us-west-2, eu-west-1, eu-central-1

last updatedJuly 1, 2026

Run m8a.large in your own account.

The benchmark runs on the same RunsOn anyone can deploy. One CloudFormation stack, your AWS account, ~10 minutes.

self-host → ← all ec2 instances