AWS CDK Reference

Appearance

v2.275.0

AcceleratorName Enum ​

Specific hardware accelerator models supported by EC2.

Defines exact accelerator models that can be required or excluded when selecting instance types.

Source
aws-ec2/lib/instance-requirements.ts
Type last changed
v2.236.0 · 23 Jan 2026

Example ​

ts
declare const vpc: ec2.Vpc;

const securityGroup = new ec2.SecurityGroup(this, 'SecurityGroup', {
  vpc,
  description: 'Security group for managed instances',
});

const miCapacityProvider = new ecs.ManagedInstancesCapacityProvider(this, 'MICapacityProvider', {
  subnets: vpc.privateSubnets,
  securityGroups: [securityGroup],
  instanceRequirements: {
    // Required: CPU and memory constraints
    vCpuCountMin: 2,
    vCpuCountMax: 8,
    memoryMin: Size.gibibytes(4),
    memoryMax: Size.gibibytes(32),

    // CPU preferences
    cpuManufacturers: [ec2.CpuManufacturer.INTEL, ec2.CpuManufacturer.AMD],
    instanceGenerations: [ec2.InstanceGeneration.CURRENT],

    // Instance type filtering
    allowedInstanceTypes: ['m5.*', 'c5.*'],

    // Performance characteristics
    burstablePerformance: ec2.BurstablePerformance.EXCLUDED,
    bareMetal: ec2.BareMetal.EXCLUDED,

    // Accelerator requirements (for ML/AI workloads)
    acceleratorTypes: [ec2.AcceleratorType.GPU],
    acceleratorManufacturers: [ec2.AcceleratorManufacturer.NVIDIA],
    acceleratorNames: [ec2.AcceleratorName.T4, ec2.AcceleratorName.V100],
    acceleratorCountMin: 1,

    // Storage requirements
    localStorage: ec2.LocalStorage.REQUIRED,
    localStorageTypes: [ec2.LocalStorageType.SSD],
    totalLocalStorageGBMin: 100,

    // Network requirements
    networkInterfaceCountMin: 2,
    networkBandwidthGbpsMin: 10,

    // Cost optimization
    onDemandMaxPricePercentageOverLowestPrice: 10,
  },
});

Members ​

MemberDescription
A100NVIDIA A100 GPU.
K80NVIDIA K80 GPU.
M60NVIDIA M60 GPU.
RADEON_PRO_V520AMD Radeon Pro V520 GPU.
T4NVIDIA T4 GPU.
V100NVIDIA V100 GPU.
VU9PXilinx VU9P FPGA.
A10GNVIDIA A10G GPU.
H100NVIDIA H100 GPU.
INFERENTIAAWS Inferentia chips.
K520NVIDIA GRID K520 GPU.
T4GNVIDIA T4G GPUs.
L40SNVIDIA L40S GPU for AI inference and graphics workloads.
L4NVIDIA L4 GPU for AI inference and graphics workloads.
GAUDI_HL_205Habana Gaudi HL-205 accelerator for deep learning training.
INFERENTIA2AWS Inferentia2 chips for high-performance ML inference.
TRAINIUMAWS Trainium chips for high-performance ML training.
TRAINIUM2AWS Trainium2 chips for high-performance ML training.
U30Xilinx U30 media transcoding accelerator for video processing.

Referenced by ​