Amazon EC2 | Cluster Instances Flashcards
Can I add an FPGA to any EC2 instance type?
Cluster Instances
Amazon EC2 | Compute
No. F1 instances comes in two instance sizes f1.2xlarge and f1.16 xlarge.
What is a Cluster Compute Instance?
Cluster Instances
Amazon EC2 | Compute
Cluster Compute Instances combine high compute resources with a high performance networking for High Performance Compute (HPC) applications and other demanding network-bound applications. Cluster Compute Instances provide similar functionality to other Amazon EC2 instances but have been specifically engineered to provide high performance networking.
Amazon EC2 cluster placement group functionality allows users to group Cluster Compute Instances in clusters – allowing applications to get the low-latency network performance necessary for tightly-coupled node-to-node communication typical of many HPC applications. Cluster Compute Instances also provide significantly increased network throughput both within the Amazon EC2 environment and to the Internet. As a result, these instances are also well suited for customer applications that need to perform network-intensive operations.
Learn more about use of this instance type for HPC applications.
What kind of network performance can I expect when I launch instances in cluster placement group?
Cluster Instances
Amazon EC2 | Compute
The bandwidth an EC2 instance can utilize in a cluster placement group depends on the instance type and its networking performance specification. Inter-instance traffic within the same region can utilize 5 Gbps for single-flow and up to 25 Gbps for multi-flow traffic. When launched in a placement group, select EC2 instances can utilize up to 10 Gbps for single-flow traffic.
What is a Cluster GPU Instance?
Cluster Instances
Amazon EC2 | Compute
Cluster GPU Instances provide general-purpose graphics processing units (GPUs) with proportionally high CPU and increased network performance for applications benefiting from highly parallelized processing that can be accelerated by GPUs using the CUDA and OpenCL programming models. Common applications include modeling and simulation, rendering and media processing.
Cluster GPU Instances give customers with HPC workloads an option beyond Cluster Compute Instances to further customize their high performance clusters in the cloud for applications that can benefit from the parallel computing power of GPUs.
Cluster GPU Instances use the same cluster placement group functionality as Cluster Compute Instances for grouping instances into clusters – allowing applications to get the low-latency, high bandwidth network performance required for tightly-coupled node-to-node communication typical of many HPC applications.
Learn more about HPC on AWS.
What is a High Memory Cluster Instance?
Cluster Instances
Amazon EC2 | Compute
High Memory Cluster Instances provide customers with large amounts of memory and CPU capabilities per instance in addition to high network capabilities. These instance types are ideal for memory intensive workloads including in-memory analytics systems, graph analysis and many science and engineering applications
High Memory Cluster Instances use the same cluster placement group functionality as Cluster Compute Instances for grouping instances into clusters – allowing applications to get the low-latency, high bandwidth network performance required for tightly-coupled node-to-node communication typical of many HPC and other network intensive applications.
Does use of Cluster Compute and Cluster GPU Instances differ from other Amazon EC2 instance types?
Cluster Instances
Amazon EC2 | Compute
Cluster Compute and Cluster GPU Instances use differs from other Amazon EC2 instance types in two ways.
First, Cluster Compute and Cluster GPU Instances use Hardware Virtual Machine (HVM) based virtualization and run only Amazon Machine Images (AMIs) based on HVM virtualization. Paravirtual Machine (PVM) based AMIs used with other Amazon EC2 instance types cannot be used with Cluster Compute or Cluster GPU Instances.
Second, in order to fully benefit from the available low latency, full bisection bandwidth between instances, Cluster Compute and Cluster GPU Instances must be launched into a cluster placement group through the Amazon EC2 API or AWS Management Console.
What is a cluster placement group?
Cluster Instances
Amazon EC2 | Compute
A cluster placement group is a logical entity that enables creating a cluster of instances by launching instances as part of a group. The cluster of instances then provides low latency connectivity between instances in the group. Cluster placement groups are created through the Amazon EC2 API or AWS Management Console.
Are all features of Amazon EC2 available for Cluster Compute and Cluster GPU Instances?
Cluster Instances
Amazon EC2 | Compute
Currently, Amazon DevPay is not available for Cluster Compute or Cluster GPU Instances.
Is there a limit on the number of Cluster Compute or Cluster GPU Instances I can use and/or the size of cluster I can create by launching Cluster Compute Instances or Cluster GPU into a cluster placement group?
Cluster Instances
Amazon EC2 | Compute
There is no limit specific for Cluster Compute Instances. For Cluster GPU Instances, you can launch 2 Instances on your own. If you need more capacity, please complete the Amazon EC2 instance request form (selecting the appropriate primary instance type).
Are there any ways to optimize the likelihood that I receive the full number of instances I request for my cluster via a cluster placement group?
Cluster Instances
Amazon EC2 | Compute
We recommend that you launch the minimum number of instances required to participate in a cluster in a single launch. For very large clusters, you should launch multiple placement groups, e.g. two placement groups of 128 instances, and combine them to create a larger, 256 instance cluster.
Can Cluster GPU and Cluster Compute Instances be launched into a single cluster placement group?
Cluster Instances
Amazon EC2 | Compute
While it may be possible to launch different cluster instance types into a single placement group, at this time we only support homogenous placement groups.
