AWS for Industries

Flexible EDA licensing in AWS with Synopsys Cloud BYOC with FlexEDA

Electronic Design Automation (EDA) tools are critical for designing silicon chips. In the cloud, pay-per-use mechanisms are the norm, but this traditionally isn’t the case for EDA licensing which can limit organizations’ ability to scale EDA on AWS. Synopsys Cloud Bring Your Own Cloud (BYOC) with FlexEDA is a groundbreaking solution that addresses this need.

Consumer demand for feature-rich, power-efficient smart devices is driving silicon design complexity to new heights, with chips containing more than 100 billion transistors already available today. To meet this growing demand, designers are employing the latest chip manufacturing technology processes, which are now entering the angstrom era. Additionally, advanced 3D packaging techniques and chiplets are becoming mainstream and are adding even more complexity that needs to be taken into account.

These requirements are driving the need for unprecedented computational growth for EDA and system analysis workloads that can peak as high as 400 percent as compared to older designs and manufacturing nodes. These workloads ebb and flow over the course of a project and design teams need scalable, pay-per-use models for EDA licenses and infrastructure in order to complete projects efficiently and economically.

AWS for EDA

Amazon Web Services (AWS) offers a broad choice of EDA-optimized compute instance types available in the Amazon Compute Cloud (EC2) and high-performance storage options that are globally available, providing an excellent choice for running EDA workloads at scale. Industry-leading customers, including Arm, NXP, Marvell, Alphawave and Astera Labs, are running semiconductor workloads on AWS today and are realizing the power of the cloud for accelerating their chip design schedules and lowering costs. In fact, Alphawave recently reported that by running on AWS, it was able to shorten its tapeout schedule by almost 1.5 months, and cloud costs were only 5 percent of the overall project cost.

AWS customers are running their most critical chip design and verification workloads in the cloud. “Semiconductor companies of all sizes are transforming their chip development on AWS, resulting in improved productivity, greater design quality, and overall lower costs,” says David Pellerin, General Manager of Industry Solutions at AWS.

However, compute and storage are not the only necessary resources for running EDA workloads at scale; EDA licenses are a major factor as well. Generally, semiconductor companies acquire a specific numbers of EDA licenses, and the number of licenses limits the number of concurrent jobs an organization can run. For example, if an organization only has 500 simulation licenses, then it can only run 500 simulations simultaneously, even when more computational resources are available. To truly use the elasticity and scalability provided by the virtually unlimited compute capacity of AWS, these companies need much more flexible licensing solutions.

Synopsys Cloud Bring Your Own Cloud (BYOC) with FlexEDA, which is available through the Synopsys Cloud BYOC solution on AWS and has a scalability and cost model very similar to that of Amazon EC2, which offers secure and resizable compute capacity for virtually any workload. Customers can scale as needed and only pay for what they use. The solution is very well suited for high-scale execution of EDA simulations and can be cost neutral at scale. For example, the cost of running simulations for 10 hours on 100 servers is the same as running for 1 hour on 1,000 servers because license and Amazon EC2 usage for both cases is 1,000 CPU hours.

Synopsys Cloud BYOC with FlexEDA on AWS

The Synopsys Cloud BYOC with FlexEDA model helps designers to overcome the dependency on EDA licenses by providing access to instant, unlimited EDA licenses. These licenses are available on a per-minute or per-hour basis through the Synopsys Cloud solution under a true pay-per-use model. With a Synopsys Cloud BYOC with FlexEDA deployment on AWS, customers can

  • run and manage their own EDA environment on AWS;
  • point their environment variable to Synopsys Cloud metering license servers in AWS;
  • run their EDA jobs with script parameters configured to as many software licenses as needed;
  • focus on designing chips faster instead of on managing license servers and overcoming scaling challenges; and
  • use the same tools, flows, and scripts they have always been using.

The reference architecture for deploying Synopsys Cloud BYOC with FlexEDA is shown below. For each customer Synopsys will implement the FlexEDA license management and metering service in an Amazon Virtual Private Cloud (VPC) provided by Synopsys and dedicated to that particular customer. The customer will enter their requirements including license and budget counts within the Synopsys Cloud Control Plane which will in turn set up the license file in the Synopsys Cloud AWS VPC for that customer, as shown in the following Figure 1.

Figure 1 Synopsys Cloud BYOC with FlexEDA Reference Architecture

The system is designed to scale license servers automatically and enable as many licenses as needed for successfully completing EDA workloads in the shortest possible amount of time. Synopsys Cloud BYOC with FlexEDA uses unique AWS capabilities to deliver this seamless EDA experience for customers who manage their own AWS environments. The patent-pending metering service uses Amazon EC2 Auto Scaling, which helps to add or remove compute capacity to meet changing demand, for auto scaling and multi-zone deployment. This capability helps the service to scale up to meet customer workload demands. For collecting telemetry, the solution uses Amazon CloudWatch —a service that monitors applications, responds to performance changes, optimizes resource use, and provides insights into operational health—and the Synopsys Cloud Portal provides a detailed analytics dashboard for customers to gain insights into EDA tool usage trends over time as shown in Figure 2.

Figure 2: Synopsys Cloud BYOC with FlexEDA dashboard

The Synopsys Cloud Portal’s budget governance capabilities help computer-aided design (CAD) managers to allocate quota for EDA tool or compute resource access by engineer, team, and product, thereby facilitating planned usage and expenditures.

“Flexible EDA licensing solutions such as Synopsys Cloud BYOC with FlexEDA help customers to scale up their semiconductor design workloads, accelerating their time to results,” says Pellerin. “AWS is proud to partner with Synopsys on this and future EDA innovations.”

This solution has been validated with Synopsys EDA workloads, scaling to thousands of cores on AWS using Amazon FSx for Lustre, which provides fully managed shared storage with the scalability and performance of the popular Lustre file system, and Amazon FSx for NetApp ONTAP, which provides fully managed shared storage built on NetApp’s popular ONTAP file system.

Astera Labs is successfully using Synopsys Cloud BYOC with FlexEDA on AWS today.

Security and Availability

Security is built into the solution architecture by design, with secrets stored in AWS Secrets Manager, which helps to centrally manage the lifecycle of secrets, and with usage data encrypted using AWS Key Management Service (AWS KMS), which lets you create, manage, and control cryptographic keys across your applications and AWS services. The solution is highly available and secure using Amazon Relational Database Service (Amazon RDS) for PostgreSQL—which helps to set up, operate, and scale a relational database in the cloud with just a few clicks—with connectivity using AWS PrivateLink, which provides private connectivity between virtual private clouds (VPCs), supported AWS services, and your on-premises networks without exposing your traffic to the public internet.

“With unmatched flexibility, highest levels of security, and on-demand software and cloud infrastructure scalability, Synopsys Cloud BYOC with FlexEDA on AWS helps design teams to focus on time to results, accelerate their schedules, and scale their cloud-based EDA environments to their specific project requirements,” says Sandeep Mehndiratta, senior vice president of Go-to-Market Sales Operations and Cloud at Synopsys. “By transforming the way that EDA resources are used, this approach centers on what designers and their designs need rather than software license availability and infrastructure limitations.”

Call to Action

Running EDA workloads on AWS has numerous benefits, including lowering costs, improving efficiency, improving design quality, and reducing time to market. However, fully using the scale of the cloud requires a scalable EDA licensing model. Synopsys has addressed this need with Synopsys Cloud BYOC with FlexEDA. To learn more and request a free evaluation of Synopsys Cloud BYOC with FlexEDA on AWS, please visit synopsys.com/cloud. To learn more about general semiconductor solutions on AWS, please visit AWS for Semiconductor and AWS Blogs for Semiconductor.

About Synopsys

Synopsys has a long history of being a global leader in electronic design automation (EDA) and semiconductor intellectual property used in the design and verification of semiconductor devices, or chips. Synopsys Cloud takes EDA to new heights, combining the availability of advanced compute and storage infrastructure with unlimited access to EDA software licenses on demand so you can focus on what you do best—designing chips, faster. Delivering cloud-native EDA tools, pre-optimized hardware platforms, an extremely flexible business model, and a modern customer experience, Synopsys has reimagined the future of chip design using the cloud.

For more information or to request a free evaluation of Synopsys Cloud BYOC with FlexEDA on AWS, please visit synopsys.com/cloud.

TAGS: ,
Ravi Poddar

Ravi Poddar

Dr. Ravi Poddar is a Senior Leader and Advisor for the Semiconductor Industry at Amazon Web Services. With over 25 years of experience in semiconductor design, he works closely with organizations in the industry to accelerate their transformation to the cloud. Ravi has held Director of Engineering positions at Pure Storage, Integrated Device Technology and Transmeta, working in the areas of ASIC and custom design methodology and physical design, both in bulk CMOS and SOI. Dr. Poddar received his BSEE with highest honors, MSEE and Ph.D. degrees from the Georgia Institute of Technology. He has over 20 publications in the areas of compute and storage optimization, parasitic extraction, circuit and device modeling, analog/mixed signal circuit design and machine learning/neural networks.