Principles of Top-Down Mixed-Signal Design
Further Benefits of Top-Down Design
ket is a result of the improved productivity and effectiveness of the design team, as described above. However, a rigorous top-down design methodology also has the bene- fit in that it allows more engineers to be effectively engaged in the development process at the same time, resulting in a further decrease in time-to-market.
As pointed out earlier, the existence of a shared executable high-level model of the sys- tem allows the test program development to be done in parallel with the block design and assembly, thereby eliminating a large contributor to the delay between when the design team and when manufacturing think the chip is ready to go. In addition, much of the final verification tasks that are needed in with a bottom-up design style are moved forward in the form of mixed-level simulation and performed in parallel by the block and top-level designers. The block developers can also get started developing models or evaluating IP while the system designer is finalizing the overall system architecture.
The improved and more formal communication that results in a rigorous top-down design methodology allows more engineers to be involved in the design process without overstressing the shared members of the team: the team lead and the top-level and sys- tem designers. There is also a natural support for hierarchy on large projects. Only two levels have been described in this paper, but a large chip can be partitioned into major sections (ex. RF, analog, digital, etc.), with overall leaders for the whole chip, as well as leaders for the individual sections.
7.5 Supports IP Reuse
Not only does the top-down design process described in this document improve the communication between the members of the design team, but when the design is being reused, it also improves communication between design teams. If the design is being sold, then the process also improved the communications between different companies: seller and buyer.
A rigorous top-down design process creates as a natural by-product a thoroughly vali- dated high-level description of the design, which is a critical enabler of IP reuse. This description is used by potential customers when evaluating the IP and by customers when integrating the IP. To see the value of having the needed model fall out of the design process as a by-product, consider the case where it does not. Using a bottom-up design process requires that the model be developed after the design was completed. This creates several barriers to the success of the IP. First, with the model not being used as an integral part of the design process it does not get much in the way of incidental testing. Substantial extra effort is required to field a high quality model, resulting in extra cost and delay. Furthermore, it is unlikely that the same quality model would be developed with an adjunct process. Second, with the model not being leveraged during the design process, the total cost of developing the model offsets any revenue from the IP, requiring higher levels of market success to break even. Finally, the model develop- ment process delays the release of the IP. This is especially troublesome as the price of IP drops dramatically as it becomes a commodity. Time-to-market is especially critical in the IP market as the price can drop by a factor of ten within a year of its release. Delay of even a month dramatically affects the total revenue of a product.
24 of 31
The Designer’s Guide Community