SiMa.ai and Synopsys
Integrated Capability to Accelerate Automotive Physical AI
At CES SiMa.ai announced the first integrated capability resulting from its strategic collaboration with Synopsys.
The joint solution provides a blueprint to accelerate architecture exploration and early virtual software development for AI-ready, next-generation automotive SoCs that support applications such as Advanced Driver Assistance Systems (ADAS) and In-vehicle-Infotainment (IVI).
As previously announced, SiMa.ai and Synopsys are collaborating to deliver machine-learning optimized, workload-verified, power-efficient SoC architectures required for software-defined vehicles with increasing levels of autonomy and more intelligent in-cabin experiences. Resulting from this strategic collaboration, the blueprint announced today enables customers to confidently jumpstart the design and validation of custom automotive AI SoCs, as well as »shift left« software development pre-silicon, helping reduce development costs, improve software quality, derisk start of production, and accelerate vehicle time-to-market.
»We are pleased with how well the two teams have worked together to quickly create a joint solution uniquely focused on unlocking physical AI capabilities for today’s software defined vehicles«, said Krishna Rangasayee, President and CEO at SiMa.ai. »Our best-in-class ML platform, combined with Synopsys industry-leading automotive-grade IP and design automation software creates a powerful foundation for innovation across OEMs in autonomous driving and in-vehicle experiences.«
»Automotive OEMs need to deliver software-defined AI-enabled vehicles faster to market to drive differentiation, which requires early power optimization and validation of the compute platform to reduce total cost of development and time to SOP«, said Ravi Subramanian, Chief Product Management Officer at Synopsys. »Our collaboration with SiMa.ai delivering an ML-enabled architecture exploration and software development blueprint supported by a comprehensive integrated suite of tools significantly jumpstarts these activities and enables our automotive customers to bring next-generation ADAS and IVI features to market faster.«
The new blueprint provides pre-integrated SoC virtual prototypes, as well as an integrated tool workflow, leveraging leading SiMa.ai and Synopsys solutions, including:
- Using the SiMa.ai MLA Performance and Power Estimator (MPPE) tool enables automotive customers to right-size their ML accelerator design for their workloads. Customers can iterate on a wide variety of accelerator configurations quickly to identify the optimal option.
- Synopsys Platform Architect is trusted by automotive companies to model automotive workloads and analyze performance, power, memory, and interconnect trade-offs at the system level before RTL, enabling informed architecture decisions early in the design cycle.
- With Synopsys’ Virtualizer Development Kit (VDK), customers can begin software
- development using a virtual SoC prototype before silicon is available, enabling full
- system bring-up within days of silicon availability and accelerating vehicle time to market by up to 12 months.
- SiMa.ai Palette SDK simplifies deployment of complex edge AI applications, supporting
- any ML workflow without compromising performance or ease of use—providing a
- complete ML stack for automotive edge solutions.
- Synopsys ZeBu emulation delivers comprehensive pre-silicon hardware/software
- performance and power validation to ensure a system architecture meets the needs of
- expected workloads.
The new joint blueprint, including supporting tool workflow, is available for early customer engagement.
