Machine types determine the physical specifications of your machines, such as the amount of memory, virtual cores, and persistent disk limits an instance will have. All machine types are currently managed by Google Compute Engine.
Machine types are divided in different classes, including:
- Standard machine types
- High CPU machine types
- High memory machine types
- Small machine types
Each machine type is billed differently. For pricing information, review the price sheet.
Available Machine Types
| Machine name | Description | Virtual CPUs1 | Memory (GB) | GCEUs2 | Max number of persistent disks (PDs)3 | Max total PD size (TB) |
|---|---|---|---|---|---|---|
n1-standard-1 |
Standard 1 CPU machine type with 1 virtual CPU and 3.75 GB of memory. | 1 | 3.75 | 2.75 | 16 | 10 |
n1-standard-2 |
Standard 2 CPU machine type with 2 virtual CPUs and 7.5 GB of memory. | 2 | 7.50 | 5.50 | ||
n1-standard-4 |
Standard 4 CPU machine type with 4 virtual CPUs and 15 GB of memory. | 4 | 15 | 11 | ||
n1-standard-8 |
Standard 8 CPU machine type with 8 virtual CPUs and 30 GB of memory. | 8 | 30 | 22 | ||
n1-standard-16 |
Standard 16 CPU machine type with 16 virtual CPUs and 60 GB of memory. | 16 | 60 | 44 | ||
n1-standard-32Beta |
Standard 32 CPU machine type with 32 virtual CPUs and 120 GB of memory. | 32 | 120 | 88 |
1For the n1 series of
machine types, a virtual CPU is implemented as a single hyperthread on a
2.6GHz Intel Xeon E5 (Sandy Bridge),
2.5GHz Intel Xeon E5 v2 (Ivy Bridge),
or 2.3 GHz Intel Xeon E5 v3 (Haswell).
This means that the n1-standard-2 machine
type will see
a whole physical core.
2 See Google Compute Engine Units.
3Persistent disk usage is charged separately from machine type pricing.
High memory machine types are ideal for tasks that require more memory relative to virtual cores. High memory machine types have 6.50GB of RAM per virtual core.
| Machine Name | Description | Virtual CPUs1 | Memory (GB) | GCEUs2 | Max number of persistent disks (PDs)3 | Max total PD size (TB) |
|---|---|---|---|---|---|---|
n1-highmem-2 |
High memory 2 CPU machine type with 2 virtual CPUs and 13 GB of memory. | 2 | 13 | 5.50 | 16 | 10 |
n1-highmem-4 |
High memory 4 CPU machine type with 4 virtual CPUs, and 26 GB of memory. | 4 | 26 | 11 | ||
n1-highmem-8 |
High memory 8 CPU machine type with 8 virtual CPUs and 52 GB of memory. | 8 | 52 | 22 | ||
n1-highmem-16 |
High memory 16 CPU machine type with 16 virtual CPUs and 104 GB of memory. | 16 | 104 | 44 | ||
n1-highmem-32Beta |
High memory 32 CPU machine type with 32 virtual CPUs and 208 GB of memory. | 32 | 208 | 88 |
1For the n1 series of
machine types, a virtual CPU is implemented as a single hyperthread on a
2.6GHz Intel Xeon E5 (Sandy Bridge),
2.5GHz Intel Xeon E5 v2 (Ivy Bridge),
or 2.3 GHz Intel Xeon E5 v3 (Haswell).
This means that the n1-standard-2 machine
type will see
a whole physical core.
2 See Google Compute Engine Units.
3Persistent disk usage is charged separately from machine type pricing.
High CPU machine types are ideal for tasks that require more virtual cores relative to memory. High CPU machine types have one virtual core for every 0.90 GB of RAM.
| Machine name | Description | Virtual CPUs1 | Memory (GB) | GCEUs2 | Max number of persistent disks (PDs)3 | Max total PD size (TB) |
|---|---|---|---|---|---|---|
n1-highcpu-2 |
High CPU machine type with 2 virtual CPUs and 1.80 GB of memory. | 2 | 1.80 | 5.50 | 16 | 10 |
n1-highcpu-4 |
High CPU machine type with 4 virtual CPUs and 3.60 GB of memory. | 4 | 3.60 | 11 | ||
n1-highcpu-8 |
High CPU machine type with 8 virtual CPUs and 7.20 GB of memory. | 8 | 7.20 | 22 | ||
n1-highcpu-16 |
High CPU machine type with 16 virtual CPUs and 14.4 GB of memory. | 16 | 14.4 | 44 | ||
n1-highcpu-32Beta |
High CPU machine type with 32 virtual CPUs and 28.8 GB of memory. | 32 | 28.8 | 88 |
1For the n1 series of
machine types, a virtual CPU is implemented as a single hyperthread on a
2.6GHz Intel Xeon E5 (Sandy Bridge),
2.5GHz Intel Xeon E5 v2 (Ivy Bridge),
or 2.3 GHz Intel Xeon E5 v3 (Haswell).
This means that the n1-standard-2 machine
type will see
a whole physical core.
2 See Google Compute Engine Units.
3Persistent disk usage is charged separately from machine type pricing.
Shared-core machine types are ideal for applications that don't require a lot of resources. Shared-core instances are more cost-effective for running small, non-resource intensive applications than standard, high- memory or high-CPU instance types.
f1-micro Bursting
f1-micro machine types offer bursting capabilities that allow instances to use additional physical CPU for short periods of time. Bursting happens automatically when your instance requires more physical CPU than originally allocated. During these spikes, your instance will opportunistically take advantage of available physical CPU in bursts. Note that bursts are not permanent and are only possible periodically.
| Machine name | Description | Virtual CPUs | Memory (GB) | GCEUs1 | Max number of persistent disks (PDs)2 | Max total PD size (TB) |
|---|---|---|---|---|---|---|
f1-micro |
Micro machine type with 1 virtual CPU, 0.60 GB of memory, backed by a shared physical core. | 1 | 0.60 | Shared CPU, not guaranteed | 4 | 3 |
g1-small |
Shared-core machine type with 1 virtual CPU, 1.70 GB of memory, backed by a shared physical core. | 1 | 1.70 | 1.38 |
1 See Google Compute Engine Units.
2 Persistent disk usage is charged separately from machine type pricing.
What are Google Compute Engine Units (GCEUs)?
GCEU (Google Compute Engine Unit) is a unit of CPU capacity that we use to describe the compute power of our instance types. We chose 2.75 GCEUs to represent the minimum power of one logical core (a hardware hyper-thread) on our Sandy Bridge or Ivy Bridge platform.
To view a list of available machine types, you can always run:
$ gcloud compute machine-types list
Using a Machine Type
You can specify the machine type when creating an instance with gcloud compute
command using the --machine-type flag:
$ gcloud compute instances create INSTANCE --machine-type MACHINE_TYPE
If you do not specify machine type when you create the instance using
gcloud compute, n1-standard-1 is used as the default.
In the API, provide your machine type as part of the request body to create an instance:
body = {
'name': NEW_INSTANCE_NAME,
'machineType': <fully-qualified-machine_type_url>,
'networkInterfaces': [{
'accessConfigs': [{
'type': 'ONE_TO_ONE_NAT',
'name': 'External NAT'
}],
'network': <fully-qualified-network_url>
}],
'disk': [{
'source': <fully-qualified-boot-disk-url>,
'boot': 'true',
}]
}