Altair > Case Studies > Rapid Chip Design in the Cloud: Annapurna Labs' Journey with Altair Accelerator

Rapid Chip Design in the Cloud: Annapurna Labs' Journey with Altair Accelerator

Altair Logo
Technology Category
  • Analytics & Modeling - Machine Learning
  • Application Infrastructure & Middleware - Data Exchange & Integration
Applicable Industries
  • Buildings
  • Construction & Infrastructure
Applicable Functions
  • Product Research & Development
Use Cases
  • Demand Planning & Forecasting
  • Time Sensitive Networking
Services
  • Cloud Planning, Design & Implementation Services
  • System Integration
About The Customer
Annapurna Labs was established as a fabless chip start-up in 2011, with a focus on bringing innovation to the fast-growing cloud infrastructure. Four years later, it was acquired by Amazon Web Services (AWS). Since then, Annapurna Labs has accelerated its innovation and developed a number of products that benefit cloud customers, including AWS Nitro technology, Inferentia custom Machine Learning chips, and AWS Graviton2 processors, based on the 64-bit Arm Neoverse architecture purpose-built cloud server.
The Challenge
Annapurna Labs, a fabless chip start-up acquired by Amazon Web Services (AWS), was facing challenges in managing workloads on dedicated Amazon Elastic Compute Cloud (EC2) instances. The team could occasionally scale up by manually adding new On-Demand instances, but the process was not automated, leading to inefficiency, forgotten unused compute resources, and either under-scaling or excessive scaling. As a chip design company, time-to-market and engineering efficiency were critical metrics for them. The team needed a solution that could add structure and efficiency to scaling AWS compute resources, shorten time to results, and change the development model to Continuous Integration.
The Solution
Annapurna Labs selected the Altair Accelerator™ job scheduler for their front-end and back-end workflows. The Accelerator's Rapid Scaling feature, developed with Annapurna Labs, automatically starts new instances only when there is demand and stops scaling up if the speed at which demand is being processed is good enough. This license-first approach to scheduling allows Accelerator to efficiently differentiate workloads waiting for licenses versus workloads waiting for hardware. Many features were added in cooperation with Annapurna Labs, including configurable selection of instance type, Spot Instance support, protection against various errors, fine control of the number of jobs that can be executed on each new instance, and many others. Rapid Scaling also understands how to select a backup instance type if the first choice is not available.
Operational Impact
  • With Rapid Scaling now part of Annapurna Labs’s chip development Continuous Integration flow, they are seeing faster incremental development and continuous regression. The company keeps tighter control on costs and benefits from a detailed view into resource usage by projects and users. The ability to understand workload speed and spin-up costs enables Rapid Scaling to avoid overshoot. This flexibility maps nicely into AWS notions of Reserved, On-demand, and Spot instances.
Quantitative Benefit
  • Annapurna Labs has reduced its cost by at least 50% with the installation of Rapid Scaling.
  • The Rapid Scaling feature allows demand to be satisfied in 10 minutes.
  • All resources are freed after they have been idle for one minute.

Case Study missing?

Start adding your own!

Register with your work email and create a new case study profile for your business.

Add New Record

Related Case Studies.

Contact us

Let's talk!
* Required
* Required
* Required
* Invalid email address
By submitting this form, you agree that IoT ONE may contact you with insights and marketing messaging.
No thanks, I don't want to receive any marketing emails from IoT ONE.
Submit

Thank you for your message!
We will contact you soon.