Leveraging IBM Cloud for electronic design automation (EDA) workloads

ByBitcoin21

Nov 4, 2023

Electronic design automation (EDA) is a market segment consisting of software, hardware and services with the goal of assisting in the definition, planning, design, implementation, verification and subsequent manufacturing of semiconductor devices (or chips). The primary providers of this service are semiconductor foundries or fabs.

While EDA solutions are not directly involved in the manufacture of chips, they play a critical role in three ways:

  1. EDA tools are used to design and validate the semiconductor manufacturing process to ensure it delivers the required performance and density.
  2. EDA tools are used to verify that a design will meet all the manufacturing process requirements. This area of focus is known as design for manufacturability (DFM).
  3. After the chip is manufactured, there is a growing requirement to monitor the device’s performance from post-manufacturing test to deployment in the field. This third application is referred to as silicon lifecycle management (SLM).

The increasing demands for computers to match the higher fidelity simulations and modeling workloads, more competition, and the need to bring products to market faster mean that EDA HPC environments are continually growing in scale. Organizations are looking at the best leveraging technologies—such as accelerators, containerization and hybrid cloud—to gain a competitive computing edge.

EDA software and integrated circuit design

Electronic design automation (EDA) software plays a pivotal role in shaping and validating cutting-edge semiconductor chips, optimizing their manufacturing processes, and ensuring that advancements in performance and density are consistently achieved with unwavering reliability.

The expenses associated with acquiring and maintaining the necessary computing environments, tools and IT expertise to operate EDA tools present a significant barrier for startups and small businesses seeking entry into this market. Simultaneously, these costs remain a crucial concern for established firms implementing EDA designs. Chip designers and manufacturers find themselves under immense pressure to usher in new chip generations that exhibit increased density, reliability, and efficiency and adhere to strict timelines—a pivotal factor for achieving success.

This challenge in integrated circuit (IC) design and manufacturing can be visualized as a triangular opportunity space, as depicted in the figure below:

EDA industry challenges

In the electronic design automation (EDA) space, design opportunities revolve around three key resources:

  1. Compute infrastructure
  2. Designers
  3. EDA licenses

These resources delineate the designer’s available opportunity space.

For design businesses, the key challenge is selecting projects that promise the highest potential for business growth and profitability. To expand these opportunities, an increase in the pool of designers, licenses or compute infrastructure is essential.

Compute infrastructure

To expand computing infrastructure on-premises, extensive planning and time are required for the purchase, installation, configuration and utilization of compute resources. Delays may occur due to compute market bottlenecks, the authorization of new data center resources, and the construction of electrical, cooling and power infrastructure. Even for large companies with substantial on-premises data centers, quickly meeting the demand for expanded data centers necessitates external assistance.

Designers

The second factor limiting realizable opportunities is the pool of designers. It is a challenge to hire designers, being highly skilled engineers, swiftly. The educational foundation required for design takes years to establish, and it often takes a year or more to effectively integrate new designers into existing design teams. This makes designers the most inelastic component on the left side of the figure, constraining business opportunities.

EDA licenses

Lastly, EDA licenses are usually governed by contracts specifying the permissible quantities and types of tools a firm can use. While large enterprises may explore enterprise licensing contracts that are virtually unlimited, they are prohibitively expensive for startups and small to medium-sized design firms.

Leveraging cloud computing to speed time to market

Firms (irrespective of size) aiming to expand their business horizons and gain a competitive edge in terms of time-to-market can strategically leverage two key elements to enhance opportunities: cloud computing and new EDA licensing.

The advent of cloud computing enables the rapid expansion of compute infrastructure by provisioning or creating new infrastructure in public clouds within minutes, in contrast to the months required for internal infrastructure development. EDA software companies have also started offering peak-licensing models, enabling design houses to utilize EDA software in the cloud under shorter terms than traditional licensing contracts.

Leveraging cloud computing and new EDA licensing models, most design houses can significantly expand their business opportunity horizons. The availability of designers remains an inelastic resource; however, firms can enhance their design productivity by harnessing the automation advantages offered by EDA software and cloud computing infrastructure provisioning.

How IBM is leading EDA

In conjunction with IBM’s deep expertise in semiconductor technology, data, and artificial intelligence (AI), our broad EDA and HPC product portfolio encompasses systems, storage, AI, grid, and scalable job management. Our award-winning storage and data transfer products—such as IBM Storage Scale, IBM Spectrum LSF and IBM Aspera—have been tightly integrated to deliver high-performance parallel storage solutions and large-scale job management across multiple clusters and computing resources.

IBM Cloud EDA infrastructure offers foundry-secure patterns and environments, supported by a single point of ownership. EDA firms can quickly derive value from secure, high-performance, user-friendly cloud solutions built on top of IBM’s industry-leading cloud storage and job management infrastructure.

In the coming months, IBM technical leaders will publish a white paper highlighting our unique capability to offer optimized IBM public cloud infrastructure for EDA workloads, serving both large and small enterprise customers.

I would be happy to discuss the challenges faced by the EDA industry and how IBM can help. Reach out to me here.

Learn more about IBM high-performance computing solutions

Source link