Unlocking Efficiency: The Power Of IP Blocks In Silicon Chip Design

The fastest, most efficient and cost-effective way to design silicon is by leveraging intellectual property (IP) blocks. This methodology reduces risk, allows a design team to focus on its own differentiation, and allows scalability. Re-using existing IP offers even more value for design teams. But not every company has embraced the approach. Here’s why you should consider it.

To optimize engineering productivity in product design, a modular design approach involves the assembly of silicon chips from pre-designed functional blocks, the IP components. These blocks range from essential elements like process design kits (PDKs that include primitive libraries), ASIC design libraries for standard-based design, and PAD libraries to protect against ESDs to more complex entities such as CPU cores (e.g., Arm/RISC-V cores), memory controllers and specialized RF design components. The methodology can be broken down into three key steps:

• IP Selection: The silicon chip design leads choose the required IP blocks or modules based on the chip’s specifications.
• IP Integration: IP blocks are integrated into the chip hierarchically once selected, adapting them to work together seamlessly at each level of hierarchy.
• Verification: Rigorous verification and simulation methods ensure the IP blocks interact correctly and function as intended within the larger chip context.

Each step involves the engineering and management teams making multifaceted and complex decisions. Keeping track of the IP selection process and its role in the project’s success is crucial for accountability, future reference, risk management, and legal/compliance. It aids quality assurance, supports knowledge transfer, and enables continuous improvement. Documenting these decisions ensures transparency, defends choices, and enhances the efficiency and effectiveness of semiconductor development projects.

Typically, IP-based engineering methodology involves three categories of IPs.

Third-party IPs: IPs procured from third-party IP vendors. The most common examples of such IPs are CPU core, memory, and accelerator. However, it is also essential to consider the basic building blocks of a chip, such as PDKs/primitive libraries and Standard Cell ASIC libraries. Considering these building blocks of IPs to track their technical (e.g., optimized power and area in the final design) and nontechnical metrics (e.g., effort required to integrate) is a valuable tool in the decision-making process for future projects.

At an organizational level, the procurement team may need to track business aspects of third-party IPs. When choosing a third-party IP, business considerations include the relationship with the IP vendor, the cost of IPs at scale, and the license to use IP for specific applications may be just as critical as the technical/operational merits of using the IP.

Corporate IPs: Semiconductor companies often acquire smaller IP companies or mandate the organic growth of specialized teams for their niche expertise. Such internal groups often act as service providers for other product teams in the companies working on SOCs. The most common and often overlooked example of this is a CAD team that handles PDKs and Model customization in medium to large semiconductor companies. Without the overhead of managing the business aspects of a third-party vendor, such units within an organization are, for all operational and technical purposes, third-party IP vendors. Such teams must act as such and track their customers, the releases used by each customer, and issues seen by each customer project effectively. Using a streamlined methodology for releases and using tracked information to plan for changes improves efficient and productive IP reuse.

Community IPs: As a good engineering practice, design teams often use trusted sources of IPs from past successful designs to build on past experience. For example, an IP such as a PLL has applications in several chips designed for various applications in an organization. But, for each application, the PLL specification/requirements are different. Rather than starting each PLL design from scratch, engineering teams can (and prefer to) customize or extend the functionality/performance of a previously designed successful PLL for a new application. This saves significant time in the design cycle. In effect, such IPs act as a template or seed design for other design teams and impact team productivity significantly. For such IPs, the organization must systematically track the IP quality, success metrics, and provenance of seed IPs to help promote IP reuse in designing high-quality products.

In conclusion, the intricate journey of silicon chip development is greatly enhanced by using IP as foundational building blocks. The modular design approach streamlines the process by integrating pre-designed functional blocks seamlessly into complex integrated circuits. Each step in this methodology involves intricate decision-making, making it essential to keep track of the IP selection process and its reasoning. This documentation ensures accountability, aids future reference, mitigates risks, and supports compliance. Whether dealing with third-party, corporate, or community IPs, the systematic traceability of technical and business aspects elevates engineering efficiency. It fosters a culture of innovation in the dynamic field of semiconductor technology.

As you navigate the world of silicon chip development, consider implementing these strategies for IP-based methodologies in your projects. Embrace documenting decisions, evaluating IPs comprehensively, and establishing streamlined procedures for effective IP management. By doing so, you enhance the efficiency of your development processes and contribute to the continuous improvement and advancement of design processes in your design workflow.