We’re excited to announce our new Compute tier, bringing you the most powerful GitHub Actions runners available for Linux workloads on both x64 and ARM64 architectures.

Built for Demanding Workloads

The Compute tier is designed for teams who need exceptional performance for resource-intensive workloads. Whether you’re running complex builds, extensive test suites, or heavy computational tasks, our compute runners deliver the power you need.

Compute tier features:

Premium CPU allocation with enhanced performance
High-speed memory for faster data access
Available for both x64 and ARM64 architectures
Compatible with all size classes (:lg, :xl, etc.)
Seamless integration with your existing GitHub Actions workflows

How It Works

Compute tier runners consume twice the concurrency slots of standard runners, delivering significantly more power while maintaining the same memory capacity. This means a compute job uses the equivalent resources of 2 regular jobs, ensuring maximum performance for your demanding workloads.

Simple to Use

Start using compute runners by prepending /compute/ to your runner label:

jobs:
  heavy-build:
    # Standard runner
    # runs-on: blaze/ubuntu-latest

    # Compute runner - same syntax, more power
    runs-on: blaze/compute/ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: make build

You can also combine compute tier with size classes:

jobs:
  intensive-tests:
    runs-on: blaze/compute/ubuntu-latest:xl # 16 vCPUs, compute-optimized
    steps:
      - uses: actions/checkout@v4
      - run: npm test

When to Use Compute Tier

Consider compute runners for:

Large Rust/C++ compilations that benefit from faster CPU cores
Parallel test suites that can utilize more processing power
Machine learning training and inference workloads
Video/image processing pipelines
Any CPU-bound task where time matters

Available Now

The compute tier is available for all accounts. Check out our documentation for detailed specifications and setup instructions.

Give it a try and experience the difference that premium compute power can make for your CI/CD pipeline.