If you fear change, ECO fill can help
By Jeff Wilson, Mentor Graphics
You know the feeling. You walk out of the grocery store with a cart full of bags, and start loading them in your car. Just as you finish, it hits you…you forgot to buy the one item you really, REALLY needed. Now you have frozen food in your backseat, so time is of the essence. How quickly can you run back in the store, get that necessary item, and be on your way home?
Design teams often feel this way as they approach tapeout, only to be confronted with engineering change orders (ECOs). One major factor—the challenge of re-filling designs. At 45 nm and below, new manufacturing requirements impacted the complexity of metal fill placement, as well as the number of fill elements in a design. Fill targets changed from targeting a basic minimum density to achieving a maximum density. In addition, density checks for density gradient were developed to ensure a smooth transition between fill densities in adjacent locations. At 20 nm and below, fill requirements had to encompass multi-patterning restrictions to ensure mask balancing, and designers had to begin adding multi-layer fill not only to the back end of line (BEOL) metal and via layers, but also to front end of line (FEOL) layers.
All of these changes in fill require sophisticated new fill types and filling strategies, because fill no longer affects just planarity, but multiple manufacturability issues. Fill now also has a direct impact on design performance. Techniques such as cell-based and MP-aware fill were developed and integrated into fill engines to provide design teams with an automated fill process that can be called from place and route tools to ensure an easy-to-use design flow that produces correct-by-construction results. However to get an accurate, optimized fill placement, designers need an environment specifically tuned for the expanding set of new checks and constraints.
Routers are designed to make millions of connections, not comply with complex fill requirements across multiple layers. Further complicating the challenges faced by a P&R system, the number of fill shapes in any of the new technologies can exceed a billion of new objects. An ECO that arrives late in the tapeout process must be handled efficiently and accurately, or the complexity of replacing fill and reconfirming timing may negatively affect file size, run time and timing closure, which can lead to a delayed tapeout delivery. To balance timing analysis against the runtime of the P&R system, the EDA industry has developed a flow in which fill shapes are kept in a separate file on disk, using an extraction tool to combine them with the layout file. This flow accounts for the timing impact of fill without slowing down the back-end flow.
An effective ECO fill strategy, then, just like our dash back into the store for the forgotten item, must be accurate and fast, concentrating only on the portion of the design affected by the ECO. By removing and replacing only the fill in that area, and re-verifying timing only in the affected area, we can reduce runtime, manage file size, and minimize timing impacts. By restricting the ECO fill operation to only the same locations where actual mask-making changes occur, we can limit the size of the region that must be evaluated for errors, edited, and refilled. This area reduction is accomplished by generating exclude regions, and clipping the fillable database to include only the area around the design ECO. This job is made easier when the designer can use the same design rule checking (DRC) signoff tool used by the foundry, to more accurately minimize these areas.
The other issue is database size. Removing fill shapes only around an ECO area requires flattening the fill hierarchy, which can make the fill database explode. To address this issue, smart ECO fill technology only flattens the minimum number of instances at the lowest level of the fill hierarchy required to remove existing fill that conflicts with the ECO design shapes. It then refills only in the areas where ECO changes occurred, rather than refilling the entire chip.
Obviously, there is a breakeven point—if the area to be refilled is too large, then the efficiencies of scale and accuracy may be lost. In general, ECO fill strategies are most efficient when the change affects less than 1% of the design area. More than that, and the runtime of the ECO fill flow may actually exceed that of a regular fill run. However, in this situation, designers must also consider whether minimizing timing impacts and mask costs offset any runtime disadvantage.
Multiple small areas of change are good candidates for ECO fill, such as changes in gate functionality that requires a localized rerouting in a limited area. Changes to an entire block indicate that it would be more efficient to simply refill the design from scratch. In these cases, a hierarchical approach would be an excellent alternative.
To make the most of these new fill techniques, not only do the tools need to provide the required functionality, but it is critical that the foundry also provides the supporting files that make the flow easy to use. As designers become familiar with ECO fill techniques, they will be able to incorporate them into their design flows to better manage late-stage design changes and maintain tapeout schedules for advanced node designs.
Jeff Wilson is a DFM Product Marketing Manager in the Calibre organization at Mentor Graphics in Wilsonville, OR. He has responsibility for the development of products that analyze and modify the layout to improve the robustness and quality of the design. Jeff previously worked at Motorola and SCS. He holds a BS in Design Engineering from Brigham Young University and an MBA from the University of Oregon. Jeff may be reached at firstname.lastname@example.org.