Intel Processors in Amazon EC2 M7i-Flex
A new Amazon Web Services (AWS) family may increase performance and save you money. The new Amazon Elastic Cloud Compute (EC2) M7i and M7i-flex instances use 4th Gen Intel Xeon Scalable processors to provide better performance per dollar for general-purpose workloads than M4, M5, and M6i. The M7i instances are the logical upgrade to the M6i, M5, and M4 instances with older Intel CPUs, while the M7i-flex instances are a new Amazon EC2 alternative that saves money on the newest technology.
The first in a three-part blog series will introduce these new instances, describe the new Intel Xeon CPU, and discuss the optimal workloads for them. Later in the series, they will examine particular workloads and utilize internal testing to demonstrate performance gains and cost reductions.
M7i Family
The newest Amazon EC2 general-purpose instances, the M7i family, balance computation, memory, network, and storage for most applications. The Amazon EC2 M7i family has nine instance sizes, from vCPU m7i.large to 192 vCPU m7i.48xlarge. The M7i series will eventually support bare-metal instances with 96 and 192 vCPUs.
The earlier M6i series had 128 vCPUs. If your workloads on m6i.32xlarge instances will soon need additional resources, you can quickly scale up to m7i.48 without spinning up new instances. As of October 16, 2023, AWS On-Demand pricing in the US East region is $6.144 per hour for m6i.32xlarge instances and $9.6768 for m7i.48xlarge instances, approximately the same per vCPU.
The M7i instances have advantages over the M6i. They enhance the amount of EBS volumes users may connect to each instance and employ the newest DDR5 RAM for greater bandwidth. You could only connect 28 network interfaces, EBS volumes, and NVMe volumes to previous-generation instances. If you added two NICs, your EBS volume limit dropped to 26. No matter the amount of network interfaces or NVMe volumes, new M7i instances can support 128 EBS volumes. Even the smallest VMs, with a 32-EBS volume limit, support greater volumes than earlier instances.
The Latest Amazon EC2 Offering Gives You Options
Amazon EC2 M7i-Flex family is a new solution for the Intel M7i instance family meant to save clients money. Some vital workloads need 24/7 resource availability, whereas others do not. Previous versions of AWS instances forced enterprises to overprovision and pay for all resources, including unused ones. Customers have to right-size every workload and instance, frequently paying for enough resources to satisfy peak demand during off-peak hours, to avoid overspending. The Amazon EC2 M7i-Flex instances save money and provide resources for unexpected load surges.
Amazon EC2 M7i-Flex promise full CPU performance 95% of the time and at least 40% in the remaining 5%. AWS may give customers a 5% discount for suffering a performance drop from re-allocating capacity to other VMs. These instances are appropriate for applications that don’t always require full computing capability but may experience demand spikes. AWS claims M7i-flex instances outperform M6i instances by 19%. Flex instances have 2–32 vCPUs and M7i-like network and storage bandwidth. They use DDR5 for higher memory bandwidth than previous versions.
Amazon EC2 M7i and Amazon EC2 M7i-Flex instances are available in several locations, and AWS is increasing them into 2023. Monitor the AWS What’s fresh page for fresh region announcements. For M7i and M7i-flex instances, AWS offers On-Demand, Reserved, Spot, and Savings Plan payment options to help clients budget.
Most savvy cloud consumers seek for processor instances first. The 4th Gen Intel Xeon Scalable processors include many built-in accelerators to boost AI, data analytics, networking, storage, and HPC performance. These accelerators utilize CPU core resources to improve power efficiency and assist organizations meet sustainability objectives.
All Amazon EC2 M7i and Amazon EC2 M7i-Flex instances use latest-generation Intel Xeon processors with the Intel AMX accelerator to increase AI job performance. When models are optimized for Intel AMX, deep learning training and inference workloads improve. Intel has updated open-source tools like TensorFlow and PyTorch with their oneAPI Deep Neural Network Library (oneDNN), making it easy to switch to the Intel AMX instruction set.
The Intel AMX accelerator is available on Amazon EC2 M7i and Amazon EC2 M7i-Flex instances allowing users to benefit from performance. Intel tested Intel AMX CPUs on several deep learning models and found that they perform 10 times better. Your company may benefit from quicker bot-driven customer chat response times, improved app predictive text, and more, depending on the AI model you select. Stay tuned for an AI-specific workload blog in this series.
Intel Data Streaming Accelerator (DSA), Intel In-Memory Analytics Accelerator (IAA), and Intel Quick Assist Technology (QAT) improve performance across workloads in the newest Intel Xeon processors. Offloading typical data transportation tasks to server components with Intel DSA boosts CPU efficiency. Intel IAA improves in-memory compression and encryption. Intel QAT offloads networking encryption and data compression to free cores and save power. Soon, bare-metal M7i instances will support these three accelerators.
Workflows
After discussing the advantages of M7i and M7i-flex instances, you may be asking which workloads are suitable for them.
If you’re performing a task on an earlier Amazon EC2 instance generation like M6i or M5, switching to M7i might improve performance and save money. Tests reveal that new M7i instances have 40% higher transaction rates for MySQL OLTP databases and 43% better CPU-intensive task performance than M6i instances. When migrating a workload to the cloud, assess its resource needs to decide whether M7i instances are right for it.
Another instance family may be better if the program is memory-intensive but CPU-light or vice versa. High CPU usage and RAM are needed for most conventional workloads such huge applications and databases, gaming servers, and video streaming apps. M7i instances can handle most public-facing and internal workloads because to their high network and storage bandwidth.
As said, M7i-flex instances are best for non-critical or variable resource workloads. Some examples?
Many web applications and online services are multi-tiered and need webpages, databases, etc. Web and application servers that host URLs and web content use less resources than the database, even when the application is resource-intensive. You might save money by hosting less resource-intensive components of your program on M7i-flex servers. This would lower the cost of capacity you normally provide to absorb demand surges. If a hot sale or viral news piece caused a sudden rise in activity, Amazon EC2 M7i-Flex instances could manage it without affecting application performance. Amazon EC2 M7i-Flex instances may host microservices, which divide up huge programs into smaller components that consume less resources.
Amazon EC2 M7i-Flex instances work nicely with virtual desktop infrastructure. Many users in your company’s divisions utilize web-based and restricted apps. Users don’t require much computational power if their desktop has adequate RAM to run these websites and apps. At the 40% CPU baseline AWS promises, Amazon EC2 M7i-Flex instances would provide enough resources for most users. You would save money and gain power users and high activity periods resources.
Less priced instances with variable resource use may assist internal databases, applications, and batch-processing workloads. Internal applications are used by fewer workers than public-facing sites and apps, even in major firms. Amazon EC2 M7i-Flex instances provide a cost-effective solution to provision staff apps for many of these workloads. Early testing suggests that new M7i-flex instances are 19% cheaper than M6i instances. They will discuss some of these workloads in depth later in this blog series and provide actual data on Amazon EC2 M7i and Amazon EC2 M7i-Flex performance gains.
Summery
Amazon EC2 M7i and Amazon EC2 M7i-Flex, the latest Amazon EC2 instance families with 4th Generation Intel Xeon Scalable processors, provide improved performance and cost savings. Host your workloads on these new instances for improved performance, processor features, and performance per money.
0 Comments