Case Study: Annapurna Labs reduces cloud compute costs by 50% and accelerates chip design with Altair's Rapid Scaling

License-first approach to scheduling enhances resource management

Annapurna Labs, a fabless chip design company now part of AWS, faced inefficient manual scaling of Amazon EC2 instances for front-end and back-end EDA workflows, leading to forgotten idle resources, under- or over-scaling, and high operational touch. To address this, Annapurna Labs selected Altair’s Accelerator™ job scheduler and worked with Altair to develop the Rapid Scaling feature to bring structure and efficiency to cloud scaling.

Altair’s Rapid Scaling uses a license-first scheduling approach in Accelerator to distinguish workloads waiting for licenses versus hardware, automatically starts and stops EC2 instances (including Spot and backup instance types), and frees idle resources after one minute so demand can be satisfied in about 10 minutes. The solution cut Annapurna Labs’ cloud compute costs by at least 50%, sped up incremental development and continuous regression, improved time-to-results and productivity, and gave tighter cost control and detailed resource-usage visibility.

Annapurna Labs

Nafea Bshara

Annapurna Labs