Chip Design Moves To The Cloud With Synopsys And Microsoft Azure

The two companies have collaborated to enable engineers to design chips using cloud resources and dynamic licensing.

Digital chips are the new lifeblood for nearly every human endeavor and currency for innovation. But design a chip is a very expensive undertaking, requiring talented engineers and tools called electronic design automation (EDA). That software typically run on servers in corporate data centers, or are deployed ad hoc on cloud instances from cloud services providers. A big reason for this is the need for absolute security and confidentiality. You certainly don’t want to risk having a chip design costing tens or hundreds of millions of dollars to slip out the door into your competitor’s hands.

That is all about to change. Designers want the flexibility of elastic cloud instances, and they want a far simpler pay-as-you-go licensing model to access the state of the art tools needed to design, engineer, and test a new product. And everyone would like to shorten the time to result from 18-24 months to less than a year.

The Synopsys EDA Cloud, on Microsoft Azure

One of the reasons EDA software has not already migrated into a SaaS model is the licensing model, which is generally inflexible and static. Synopsys has come up with a new model wherein a design team can rent the software by the hour, across an increasingly rich portfolio of tools. The design team can now begin to deploy that software manually on a cloud in a Bring-your-own-cloud model, or can use pre-installed and configured EDA tools as an on-demand service from Microsoft Azure. This complements the existing Synopsys ZeBu Cloud platform, which delivers turn-key emulation for bring-up, performance validation, power analysis, and system validation for IP and SoCs.

The Synopsys-managed SaaS offering on Azure is configured for a select set of tools today, but we expect it to eventually offer practically all of Synopsys’ EDA portfolio, including the DSO.ai platform that requires substantial hardware clusters and accelerators. Therein lies the other magic ingredient to complement the on-demand licensing: elastic compute services configure to support specific tasks and tools. The user doesn’t have to worry about what hardware is best, nor how to afford, install and configure it. All that “dirty work”, an expensive nuisance to design teams, simply vanishes.

Conclusions

While there is a lot more to unpack here soon, this announcement is an important milestone in the ongoing advances of cloud computing, and will be welcome news to chip designers around the world.

Disclosures: This article expresses the opinions of the author, and is not to be taken as advice to purchase from nor invest in the companies mentioned. My firm, Cambrian AI Research, is fortunate to have many, if not most, semiconductor firms as our clients, including Blaize, Cerebras, Esperanto, Graphcore, IBM, Intel, NVIDIA, Qualcomm Technologies, Synopsys, and Tenstorrent. We have no investment positions in any of the companies mentioned in this article and do not plan to initiate any in the near future. For more information, please visit our website at https://cambrian-AI.com.