What is Triple Shift Left?


The term “shift left” originated in the software industry and refers to finding and fixing bugs early in the development process rather than catching them during post-release testing. IBM found that fixing a software defect during the development phase is 100 times less expensive than fixing it after release. And the cost stakes are even higher with semiconductors.

For more than 30 years, Synopsys has been helping system-on-chip (SoC) designers keep up with growing complexity, shrinking geometries, increasing hardware/software interdependency, and aggressive schedules through a strategy we call "Triple Shift Left." It's dedicated to solving the high-value problems of functional safety, security, and reliability for the automotive industry, where requirements are evolving to support mission-critical applications for embedded vision, V2X communications, autonomous driving, and more.

Triple Shift Left transforms a traditionally serial development process into a parallel one. It enables designers to find mistakes earlier in the design process where they are not only cheaper to fix, but also where OEMs can have visibility into the earliest parts of the design through virtual prototypes. With a Triple Shift Left strategy, design teams can build functional safety and reliability into their SoCs, faster; identify software problems up to 18 months earlier, before hardware is available; and build security and quality into their software during development and testing, and across the supply chain.

How Does Triple Shift Left Work?

Shift Left I: Smarter, Safer Automotive SoC Design
Rather than starting from scratch, SoC designers save staff-years of effort by front-loading their designs with safe, secure, reliable, and reusable building blocks (known as IP). Synopsys DesignWare IP for automotive includes ISO 26262-compliant logic, interfaces, safety islands, and embedded vision processors. It ensures various levels of ASIL (A, B, C, and D) compliance and meets strict AEC-Q100 reliability standards. Using Synopsys auto-grade IP, designers can build dedicated functions on silicon, reduce integration risk, and shift even further left to design functional safety and reliability into their SoCs from the start.

Shift Left II: Parallel Software and Hardware Development
Virtual prototyping allows hardware and software to be developed concurrently. As a result, software development for an ECU can be completed up to 18 months before actual silicon manufacture. Virtual prototyping also facilitates collaboration across the automotive supply chain, enabling early software development and defining chip architectures to ultimately increase the quality of results and reduce time to production. Synopsys accelerates the development and deployment of virtual prototypes.  

Shift Left III: Comprehensive Automotive Software Testing
Further leveraging virtual models, Synopsys helps customers create test platforms for virtual vehicle development. This enables extensive verification through static security testing, software composition analysis, interactive security testing, and fuzz testing. Scaling and derivative testing also become possible, allowing for experimentation with countless potential hardware-software configurations to reduce verification costs for low-volume derivatives. This regression forming, or automated testing, increases overall coverage and accelerates test cycles for applications, power electronics, wire harness simulations, and a variety of other specific applications. With Synopsys software integrity solutions, customers can build security, quality, and compliance into their automotive software lifecycle.

Can Triple Shift Left Enable Autonomous Driving?


Autonomous cars are the future of the automotive industry. They’ll require exponentially more content for both electronic hardware and software. The development of next-generation automotive chips is already forcing the convergence of conflicting goals like increased reliability, reduced costs, and shorter development cycles. Going driverless will require a transformation of today’s automotive design process. This new design process will require two distinct changes:

Change #1: Dramatically increase automotive reliability
The present 10% failure rate of today’s automotive process is too high to support self-driving cars. If the industry expects to proliferate fully autonomous vehicles, it must aim for an audacious goal: zero defects. And while that goal will never be achieved fully, adopting a Triple Shift Left strategy will infinitely drive the failure rate toward zero.

One way to increase reliability rates is to add redundancy. By replicating functionality as a backup, failures of individual components will not cause a failure of the entire system. The complexity of increased components makes the system more robust and less likely to fail, thereby increasing the FIT (failure in time) rate.

Change #2: Combine functionality to reduce BOM costs
While adding redundancy increases reliability, doing so through discrete components has the potential to balloon bill-of-material (BOM) costs. It’s estimated that adding ISO 26262-compliant ADAS technologies adds $10,000 to a new car’s BOM, while full autonomy will add a whopping $120,000. The only way to deliver superior next-gen products at consumer prices is to add more functionality into fewer components. New electronics will require a new strategy like Triple Shift Left.

Get Started with Triple Shift Left