TetraMem Delivers RISC-V AI Accelerator Tape-Out in Record Time on Synopsys Cloud

Anuj Pant

Mar 27, 2024 / 3 min read

Founded in Silicon Valley in 2018, TetraMem is ready to change the landscape for AI computing by addressing one of its most vexing challenges: the need to deliver massive compute processing while keeping its power consumption in check. With its in-memory computing (IMC) technology for efficient AI applications, TetraMem is the only company producing a high-bit-density, multi-level RRAM (otherwise known as a computing memristor) accelerator in commercial foundries.

Comprised of a global team spread across different continents, TetraMem needed an efficient way to collaborate to develop its hardware solutions, as well as scalable and flexible design and verification resources. The company found its answer in cloud-based EDA solutions, namely through Synopsys Cloud. Providing cloud-native EDA tools and pre-optimized hardware platforms, Synopsys Cloud supports chip design from end to end, with advantages including end-to-end license management automation, robust security, and spot instances for high-memory EDA workloads.

“TetraMem’s analog RRAM-based in-memory computing technology, inspired by the human brain, changes the paradigm of how AI computations are performed, launching new possibilities for AI everywhere," said Dr. Glenn (Ning) Ge, CEO and co-founder of TetraMem. "Designing for a radically different system requires reliable, robust, and flexible engineering tools and collaboration across global working sites." 

cloud eda tools

Synopsys Cloud has played a pivotal role in helping us accomplish our mission by providing seamless access to a highly secure and complete SoC design environment. With scalable access to EDA software, pre-configured end-to-end flows, and infrastructure resources, we can tape out quickly and efficiently."

Dr. Glenn (Ning) Ge, CEO and Co-Founder, TetraMem


Faster Time to Results and Elastic Compute Resources

The TetraMem team knew they needed smooth integration and verification of its complex AI accelerator’s analog in-memory computing technology and its digital RISC-V processor. Since project demands on EDA resources can fluctuate depending on the design stage, the team recognized that having scalable, elastic, on-demand access to these resources would be advantageous. Aggressive time-to-market targets also weighed on the team.

Synopsys Cloud proved to be the answer to all these challenges, enabling TetraMem to experience:

  • A jump start on hardware development through a pre-configured, end-to-end analog design flow
  • An intuitive, user-friendly browser-based platform for chip design
  • Faster time to results through on-demand, pay-per-use access to unlimited EDA resources without licensing constraints
  • Instant availability of scalable and advanced compute and storage resources

“We were able to achieve a very fast infrastructure setup on the Synopsys Cloud EDA environment within days,” said Wenbo Yin, VP of IC Design at TetraMem. “The vast selection of EDA tools and IP available on the cloud enabled us to start the design, verification, and backend flow very quickly. The flexibility of using as many licenses as we needed enabled us to obtain fast turnarounds on simulation, verification, and backend flow, which significantly reduced the engineering hours spent on those compute-intensive tasks. The ease of global access provided a single environment for our global R&D team with seamless access.”

Notes David George, TetraMem’s head of global operations: “We are creating the world’s most efficient AI accelerator chip, built around a RISC-V architecture and powered by our novel analog in-memory computing technology. Developing and integrating these elements requires a well-engineered platform to support an iterative and flexible design process in order to meet performance objectives with quality. With Synopsys Cloud, we are able to access a complete, pre-configured design flow that enables our global R&D teams to collaborate efficiently. This improved productivity for our entire team, and helped us architect a robust design for a first-of-its-kind technology.”

In addition to the technologies available through Synopsys Cloud, the TetraMem team also tapped into Synopsys’ worldwide support team. “One of our challenges from the start was bringing global team members up to speed to meet our product specifications and time-to-market goals,” said Charles Wei, the company’s director of SoC design. “The Synopsys sales, technical support, and R&D teams worked together to help us become familiar with the platform. Any EDA tool upgrades or new tool purchases are delivered immediately without worrying about CAD installation, global environment setup, version compatibility, etc. We trust that using Synopsys Cloud and collaborating with Synopsys for the development of our current and future SoC designs is the right choice.”

Flexible Licensing and Deployment Options

Synopsys Cloud features a FlexEDA business model with two licensing options. The pay-per-use (PPU) option is a usage-based licensing approach offering by-the-minute pricing for EDA tools, while a cloud subscription license (CSL) option is term-based and requires upfront payment. Both PPU and CSL options use cloud credits. Synopsys Cloud can be deployed through Bring-Your-Own-Cloud (BYOC) or Software-as-a-Service (SaaS) models.

Read other Synopsys Cloud customer success stories here.

Continue Reading