Welcome to Issue # 27 of AWS Graviton Weekly, which will be focused on sharing everything that happened in the past week related to AWS Silicon: from March 3rd, 2023 to March 10th, 2023.
Enjoy the content of this week.
The recommended readings for this week?3 things
Read how Wealthfront is saving up to 20% and 8 hours of Data processing across 500 Apache Spark jobs on Graviton
If you want to save big in your cloud bills (not just AWS), you must check outVantage. They recently raised a $21 Million Series A round from amazing investors because the product is incredible, the team is very amazing and they are cash positive (a very rare and elusive thing in private tech companies )
And if you are using Snowflake, make sure to read the announcement they made about how they see customers using Graviton and themselves as well
We are excited to announce support for the Amazon Graviton2 instance family in four additional regions. Supported instance types include compute optimized (C6g), general purpose (M6g), and memory optimized (R6g) instances.
Support for C6g, M6g, and R6g is available in the Asia Pacific (Hyderabad), Asia Pacific (Jakarta), Asia Pacific (Melbourne), and Europe (Milan) regions.
Amazon OpenSearch Service Graviton2 instances support OpenSearch versions and Elasticsearch versions 7.9 and above. With Amazon OpenSearch Service, Graviton-based instances provide up to 30% better price-performance than comparable x86-based Amazon Elastic Compute Cloud instances. Further savings are available through reserved instance (RI) pricing for these instances.
Some of my favorite parts of the announcement is precisely this:
Here are some high level updates:
- Vantage helps customers manage over $1B dollars in annualized cloud costs.
- Over 300 customers trust Vantage as their FinOps platform including organizations like Square, NASA, Buzzfeed, PBS, Panther Labs, Ripple, Compass, Rippling, Joybird, Starburst, PlanetScale and Barstool Sports
- Vantage has gone from supporting only AWS to 10 natively supported providers: AWS, Azure, Google Cloud, Snowflake, Databricks, Datadog, MongoDB Atlas, New Relic, Fastly and Kubernetes.
- There are now over 10,000 connected infrastructure accounts to Vantage across all of the providers mentioned above.
Two of my favorite features of the product are precisely Cost Recommendations and Autopilot, which could be perfect to save big in your cloud bills.
Amazon Web Services (AWS) has expanded the number of virtual machines based on the latest generation of its Graviton processors to include instances that support DDR-5 memory.
The latest Amazon EC2 instance types, dubbed M7g and R7g, that are now generally available are based on the AWS Graviton3 processors developed using an architecture created by Arm. Previously, these instances were only available as part of a technology preview program.
This report focuses on how to tune a Spark application to run on a cluster of instances. We define the concepts for the cluster/Spark parameters, and explain how to configure them given a specific set of resources.
We use a K-Means Machine Learning algorithm as a case study to analyze and tune the parameters to achieve the required performance while optimally using the available resources.
When this report was written Graviton3 only offered compute-optimized instances (C7g).
In this Guidance, we help customers simplify the deployment of an ETH2 validator on Rocket Pool, powered by AWS Graviton-based instances. Rocket Pool is one of many decentralized staking pools that supports the Ethereum beacon chain.
Validators who wish to earn rewards for staking may use Rocket Pool to create their own validation nodes. This Guidance uses AWS Cloud Development Kit (AWS CDK) and AWS Graviton-based instances to deploy ETH2 validators on the AWS Cloud with minimal setup time.
DeepSparse is an inference runtime with GPU-class performance on CPUs. With DeepSparse, deep learning deployments benefit from the scalability and flexibility of software running on commodity hardware while meeting the performance demands of production, enabling you to simplify your operations and reduce your infrastructure costs. DeepSparse on AWS Graviton enables you to run inference on the most cost-effective EC2 instances on AWS.
And by the way, it doesn't impact all customers. It only impacts some customers. And it's counterintuitive to the CPU enhancements on Graviton 2, and I'll explain that with warehouse scheduling service what this makes, is it makes your processing of your queries more efficiently, efficient when you're scheduling.
So, if you, in your warehouse, you just have one massive warehouse of data and you're running just really, really large queries and a few of them, okay, warehouse scheduling service has no impact on your performance, right?
But if you're a customer where you have a lot of data, but hundreds and hundreds of small queries you're running, that's going to have a dramatic impact on your performance in which you're using.
Now, switching it to CPU improvement, with the CPU improvement we're seeing with Graviton2 AWS’ proprietary chip, we see it really varies by customer as example, Snowflake, we saw 5% performance improvement.
We see other companies with zero and yet we've seen a few smaller ones, 30% improvement depends upon what you're doing.
If you are running really, really large queries that are really compute intensive; you're going to see a huge impact.
But if you're running a lot of small queries where there's still a lot of read, right, that has to go back to memory, you don't see the same. You don't get the full benefit of the CPU because you're IO bound.
So, it's not as easy to predict until you really put it into production, but we feel that every year we'll have these 5% revenue headwinds associated with these things.
CIO’s are now being pulled into yet another discussion: Cloud responsibility and sustainability.
In Episode 5 of CloudBrew, CoreStack's NextGen Cloud Governance podcast, we call on the experts to discuss what real sustainability is and why it’s critical in cloud. What’s AWS doing for sustainability? What about CoreStack?
It's often said that “Porting to Arm is boring,” but how easy is it, really? We'll demonstrate top machine learning frameworks, HPC applications, and tools for data science on the NVIDIA Arm HPC DevKit. The DevKit is an on-ramp platform for NVIDIA Grace CPU that incorporates dual NVIDIA A100 GPUs, an NVIDIA BlueField DPU, and an 80-core Ampere Altra Arm CPU, in a standards-compliant Arm server. We'll walk through the complete installation process for key applications and their dependencies including codes like TensorFlow, OpenRADIOSS, WRF, GROMACS, BWA-MEM2, and Jupyter Notebook. We'll also show how codes incorporating x86 AVX and SSE SIMD instructions can be trivially ported to Arm with freely available tools. We'll conclude with a general guide to porting to NVIDIA Grace and links to downloadable resources and tutorials that fully replicate our demonstrations. This session is a strong starting point for anyone targeting NVIDIA Grace Hopper or the NVIDIA Grace CPU Superchips.
¿Conoces las instancias Spot y Graviton de AWS y sus mejoras para este 2023? Los servicios de infraestructura moderna que os ayudarán a reducir vuestra factura de cómputo hasta un 70%. Miles de clientes en AWS ahorran tiempo y dinero usándolas en Servicios web, cargas en contenedores, Big data analytics (Hadoop, Spark...), procesos en Batch, CI/CD o Bases de datos como MySQL, PostgreSQL, MariaDB... Además, estos servicios se aplican en entornos productivos.