Issue # 82: April 5, 2024 to April 12, 2024

Hey Reader.

Welcome to Issue # 82 of AWS Graviton Weekly, which will be focused on sharing everything that happened in the past week related to AWS Silicon: from April 5, 2024, to April 12, 2024.

This week has been very quiet related to AWS Silicon because most headlines were occupied by Google Next, especially the new Arm-based CPU called Axion, which looks very interesting and promising.

But one thing I noticed was that more and more times we will see AWS Silicon present in Amazon's shareholders' letters by Andy Jassy

Arm is becoming ubiquitous to cloud-focused CPUs. The future looks bright for them and the whole ecosystem.

This is why I'm investing a lot of time creating an amazing Notion-based content library focused on ARM custom chips on the cloud (Graviton, AmpereOne, Trainium, Inferentia, Cobalt, and now Axion).

It hasn't been easy, dedicating more than 780 hours to curate, review, organize, and tag more than 2,500 resources related to ARM custom chips.

But unfortunately, it's not ready yet.

A sneak peek?

Stay tuned for the next week and you will be the first to know.

Back to biz

Tool of the week: CAST AI​

If you are using Kubernetes on the cloud (AWS, GCP, or Azure), according to a new report from CAST AI, perhaps you are overspending on Kubernetes; and right now, every penny overspent; it's a penny that you can't use to what truly matters: growing your company and services.

CAST AI is the "magic wand" your company needs to save up to 60 percent of your Kubernetes-related cloud bills. Believe me: you will thank me later.

NEWS

• ​CEO Andy Jassy’s 2023 Letter to Shareholders​
• ​AWS Lambda functions now scale up to 12X faster in the AWS GovCloud (US) Regions​
• ​Announcing availability of AWS Outposts in the Kingdom of Saudi Arabia​
• ​Amazon EC2 C7i instances now available in Europe (Paris)​​
  ​
  ​

ARTICLES AND TUTORIALS

• ​Introducing Google Axion Processors, our new Arm-based CPUs​​
  ​
  ​
• ​How Automation Reduces Large Language Model Costs, by CAST AI​

​

• ​Tackle complex reasoning tasks with Mistral Large, now available on Amazon Bedrock​​
  ​
  ​
• ​How Snap Inc. secures its services with Amazon EKS​​
  ​

Newsletter recommendation of the week

Machine Learning, RecSys, LLMs, Engineering. Eugene Yan Building machine learning systems @ Amazon. Writing about ML, RecSys, and LLMs @ eugeneyan.com. Join 6,000+ readers! Subscribe​

SLIDES, AUDIO AND VIDEO

• [PODCAST] JIT Assembly to Build Exascale AI Infrastructure | Neural Search Talks: Ash Vardanian (Unum)

EVENTS

​Cost Optimization for Compute Workshop
(Monday 15th April 08:00 - 11:00 GMT-5)​

At AWS, saving customers money is part of our DNA. It was a founding principle of AWS back in 2006 and continues to shape everything we do, from how we develop products to how we build greater efficiency into our operations. AWS is relentless in inventing new ways to increase performance while optimizing costs and today offers the broadest and deepest functionality for computing.

Speakers:

• ​Christian Melendez (AWS)
• ​Syl Taylor (AWS)
• ​Louis Demerliac (AWS)

​Berlin AWS User Group Meetup - April 2024​

Are you ready for an exciting evening of learning and networking? We're delighted to invite you to our upcoming event on April 16th, for another great AWS User Group session in Berlin. The session is sponsored by Rackspace Technology and will feature a great talk, followed by an AMA.

19:00 - 19:45 - Cristian Măgherușan-Stanciu // RDS and ElastiCache Optimization Strategies

In this session, Cristian will share how you can cut costs and reduce the carbon footprint of your Amazon RDS and ElastiCache clusters without giving up on performance.

We'll cover everything from choosing the right compute options — like selecting the best instance types, including Graviton where supported, good metrics for rightsizing, etc. — to picking the best EBS volume types for database storage.

We'll also cover scenarios where it makes sense to use the available Serverless options for RDS and Elasticache.

​

​

From the ARM Ecosystem

• ​Accelerate AI Inference with Google Cloud TPUs and GPUs​​​
  ​​
• ​What’s new with Google Cloud’s AI Hypercomputer architecture​​​
  ​​

​

​​

​