List of available Amazon Web Services US instances

The list below shows available AWS instances that you can choose and specify as a value for the sbg:AWSInstanceType hint. All instances listed below are available in both US East (N.Virginia) and US West (Oregon) regions that can be selected as project locations on the Platform.

See the AWS page on instance types for details on pricing.

List of instances with ephemeral storage

Name	Cores	RAM [GB]	Storage [GB]	Network Performance (Gbps)
i2.xlarge	4	30.5	800	Moderate
i2.2xlarge	8	61	1600	Moderate
i2.4xlarge	16	122	3200	High
i2.8xlarge	32	244	6400	10 Gigabit
i3.large	2	15.25	1 x 0.475 NVMe SSD	Up to 10
i3.xlarge	4	30.5	1 x 0.95 NVMe SSD	Up to 10
i3.2xlarge	8	61	1 x 1.9 NVMe SSD	Up to 10
i3.4xlarge	16	122	2 x 1.9 NVMe SSD	Up to 10
i3.8xlarge	32	244	4 x 1.9 NVMe SSD	10
i3.16xlarge	64	488	8 x 1.9 NVMe SSD	25
i3en.large	2	16	1 x 1.25 NVMe SSD	Up to 25
i3en.xlarge	4	32	1 x 2.5 NVMe SSD	Up to 25
i3en.2xlarge	8	64	2 x 2.5 NVMe SSD	Up to 25
i3en.3xlarge	12	96	1 x 7.5 NVMe SSD	Up to 25
i3en.6xlarge	24	192	2 x 7.5 NVMe SSD	25
i3en.12xlarge	48	384	4 x 7.5 NVMe SSD	50
i3en.24xlarge	96	768	8 x 7.5 NVMe SSD	100
x1.16xlarge	64	976	1920	10
x1.32xlarge	128	1952	1920	25

List of instances with variable attached EBS storage

The default attached storage size is 1 TB, but it can be changed to anything between 2GB and 4 TB.
Learn more from our EBS Customization Documentation

Name	Cores	RAM [GB]	Auto-scheduling*	EBS Bandwidth (Mbps)	Network Performance (Gbps)
c4.large	2	3.75	Yes	500	Moderate
c4.xlarge	4	7.5	Yes	750	High
c4.2xlarge	8	15	Yes	1000	High
c4.4xlarge	16	30	Yes	2000	High
c4.8xlarge	36	60	Yes	4000	10 Gigabit
c5.large	2	4	Yes	Up to 4,750	Up to 10
c5.xlarge	4	8	Yes	Up to 4,750	Up to 10
c5.2xlarge	8	16	Yes	Up to 4,750	Up to 10
c5.4xlarge	16	32	Yes	4,750	Up to 10
c5.9xlarge	36	72	Yes	9,500	10
c5.12xlarge	48	96	Yes	9,500	12
c5.18xlarge	72	144	No	19,000	25
c5.24xlarge	96	192	No	19,000	25
m4.large	2	8	Yes	450	Moderate
m4.xlarge	4	16	Yes	750	High
m4.2xlarge	8	32	Yes	1000	High
m4.4xlarge	16	64	Yes	2000	High
m4.10xlarge	40	160	Yes	4000	10 Gigabit
m4.16xlarge	64	256	No	10000	25 Gigabit
m5.large	2	8	Yes	Up to 4,750	Up to 10
m5.xlarge	4	16	Yes	Up to 4,750	Up to 10
m5.2xlarge	8	32	Yes	Up to 4,750	Up to 10
m5.4xlarge	16	64	Yes	4,750	Up to 10
m5.8xlarge	32	128	Yes	6,800	10
m5.12xlarge	48	192	Yes	9,500	12
m5.16xlarge	64	256	No	13,600	20
m5.24xlarge	96	384	No	19,000	25
r4.large	2	15.25	Yes		Up to 10
r4.xlarge	4	30.5	Yes		Up to 10
r4.2xlarge	8	61	Yes		Up to 10
r4.4xlarge	16	122	Yes		Up to 10
r4.8xlarge	32	244	Yes		10
r4.16xlarge	64	488	No		25
r5.large	2	16	Yes	Up to 4,750	Up to 10
r5.xlarge	4	32	Yes	Up to 4,750	Up to 10
r5.2xlarge	8	64	Yes	Up to 4,750	Up to 10
r5.4xlarge	16	128	Yes	4,750	Up to 10
r5.8xlarge	32	256	Yes	6,800	10
r5.12xlarge	48	384	Yes	9,500	12
r5.16xlarge	64	512	No	13,600	20
r5.24xlarge	96	768	No	19,000	25

* Instances labelled with Yes in the auto-scheduling column are the instances that can be selected for task execution automatically based on the defined CPU and memory requirements. To be able to use instances that are not available for automatic scheduling, you must set the instance type explicitly using the sbg:AWSInstanceType hint.

GPU Instances

The Platform also supports the following powerful, scalable GPU instances that deliver high performance compute in the cloud. Designed for general-purpose GPU compute applications using CUDA and OpenCL, these instances are ideally suited for machine learning, molecular modeling, genomics, rendering, and other workloads requiring massive parallel floating point processing power.

Name	GPUs	vCPUs	RAM (GiB)	EBS Bandwidth	Network Performance
p2.xlarge	1	4	61		High
p2.8xlarge	8	32	488		10 Gbps
p2.16xlarge	16	64	732		20 Gbps
p3.2xlarge	1 Tesla v100	8	61	1.5 Gbps	Up to 10 Gbps
p3.8xlarge	4 Tesla v100	32	244	7 Gbps	10 Gbps
p3.16xlarge	8 Tesla v100	64	488	14 Gbps	25 Gbps
g4dn.xlarge	1	4	16	Up to 3.5 Gbps	Up to 25 Gbps
g4dn.2xlarge	1	8	32	Up to 3.5 Gbps	Up to 25 Gbps
g4dn.4xlarge	1	16	64	4.75 Gbps	Up to 25 Gbps
g4dn.8xlarge	1	32	128	9.5 Gbps	50 Gbps
g4dn.16xlarge	1	64	256	9.5 Gbps	50 Gbps
g4dn.12xlarge	4	48	192	9.5 Gbps	50 Gbps

Creating Docker images containing tools that are run on GPU instances is similar to the process of creating Docker images with tools that are designed for CPU instances. The only major difference is that GPU tools have additional requirements for interaction with GPUs, which can be either OpenCL or CUDA. NVIDIA drivers come preinstalled and optimized according to the Amazon best practice for the specific instance family and are accessible from the Docker container. It is recommended to use one of Docker images provided by NVIDIA as the base image. For tools that require CUDA, the list of supported images are available at https://hub.docker.com/r/nvidia/cuda/, and for tools that are based on OpenCL at https://hub.docker.com/r/nvidia/opencl. The rest of the procedure for creating and uploading a Docker image is the same as for tools designed to run on CPU instances. In case you have any problems with the setup, please contact our Support Team.

When creating a Docker image containing GPU tools, it should be taken into account that older binaries are usually built for older GPU architectures and might not work on newer GPUs. If that is the case, these binaries can’t be used, and new ones should be built from source code.