Intelligent DTCO (iDTCO) for next generation logic path-finding

Prior to 90nm the main contributor to yield loss was particle contamination. Random particle defects were directly proportional to particle density and chip size. To maximize yield, manufacturing was done in clean rooms where particles were reduced to a minimum. Yield was then directly proportional to the chip size. At 90nm and below, systematic defects play major roles in yield. Two designs of the same size can result in very different yield. This paradigm shift is caused by the printability problems that are inherent to process technologies operating in the sub-wavelength regime. The technology for printing patterns on silicon has not kept up with Moore's law. The wavelength of lithography equipment is as much as 6 times larger than the silicon features. This results in greater variation on silicon, which directly affects yield. How the design is implemented is becoming as important as how small you make it. Even at the standard cell level different solutions can demonstrate increased sensitivity to particles, short, opens, gate leakage and other yield issue. We present in this paper a methodology for grading how well standard cells will print on silicon. Using standard cell layouts, we predict a silicon image under different process conditions and take CD measurements on these images. These measurements are converted into a printability factor for each cell. This printability factor is used to grade cells and identify which cells have the largest impact on printability and which should be optimized. To help with this optimization the measurements are also used to mark printability hot spots in the cell layouts.

Standard cells are the most critical and reusable elements to build up the whole chip, therefore foundry has to fully qualify the standard cell libraries to ensure their high quality when releasing to the customers for the chip design. To prevent pattern dependent lithographic difficulty in manufacturing is one target of standard cell qualification and becomes mandatory especially in advanced nodes due to tighter design rules and smaller design size. To identify a lithographic problematic standard cell, we have to take its surroundings within the optical diameter range into consideration because lithographic effects are intrinsically context-dependent. One critical step is to imitate standard cells placements in real designs and consider some important factors like VIA location as it impacts the mask shape directly. When the placement is completed, lithographic simulation is performed by LFD (Litho Friendly Design) to highlight risky locations. Every standard cell has to occur enough number of times to make sure the statistics of possibility of being a problem is reliable. The final statistics will instruct engineers on how to handle the problematic standard cells, either standard cell layouts have to be optimized or building a pattern database to prevent the abutments of particular standard cell combinations.

This paper presents a novel timing-driven global routing algorithm for standard cell (SC) layout design based on a critical network concept. The essence of this algorithm is different from that of the typical existing methods, such as the nets-based method and the critical-path-based method. The timing optimization strategy presented in this algorithm is more optimal than the typical existing ones, which makes it possible to reduce the delay in an overall survey. This algorithm has been implemented and tested by MCNC benchmark circuits in this paper. We compared the experimental results between our algorithm and the existing ones. The experimental results show that this algorithm is able to control the delay of the circuit efficiently, and causes little negative effect on other optimizing objectives.

Process-induced mechanical stress is used to enhance carrier transport and achieve higher drive currents in current complementary metal-oxide-semiconductor technologies. This paper explores how to fully exploit the layout dependence of stress enhancement and proposes a circuit-level, block-based, stress-enhanced optimization algorithm that uses stress-optimized layouts in conjunction with dual- <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> assignment to achieve optimal power-performance tradeoffs. We begin by studying how channel stress and drive current depend on layout parameters such as active area length and contact placement, while considering all layout-dependent sources of mechanical stress in a 65 nm industrial process. We then investigate the three main layout properties that impact mechanical stress in this process and discuss how to improve stress-based performance enhancement in standard cell libraries. While varying the stress-altering layout properties of a number of standard cells in a 65 nm industrial library, we show that ?dual-stress? standard cell layouts (analogous to ?dual- <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> ?) can be designed to achieve drive current differences up to ~ 14% while incurring less than half the leakage penalty of dual- <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> . Therefore, when the flexibility of ?dual-stress? assignment is combined with dual- <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> assignment (within the proposed joint optimization framework), simulation results for a set of benchmark circuits show that leakage is reduced by ~ 24% on average, for iso-delay, when compared to dual- <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> assignment. Since mobility enhancement does not incur the exponential leakage penalty associated with <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> assignment, our optimization technique is ideal for leakage power reduction. However, our framework can also be used to achieve higher performance circuits for iso-leakage and our joint optimization framework can be used to reduce delay on average by ~ 5%. In both cases, the proposed method only incurs a small area penalty (<0.5%).

We propose a flow for routing nets within a standard cell that 1) generates candidate routes for point-to-point segments; 2) finds conflicts (electrical shorts and geometric design rule violations) between candidate routes; and 3) solves a SAT instance producing a legal and complete routing for all nets in the standard cell. This approach enables routing automation for cutting-edge process technology nodes. We present how to make this technique more effective by introducing pruning techniques to reduce the work required in all three steps. We also show how we can further optimize routing quality within the SAT formulation through the use of successively more stringent constraints. Recent improvements in the speed of SAT solvers make such a formulation practical for even complex standard cells. A routing tool based on our SAT formulation is currently being used to route real industrial standard cell layouts. It demonstrates acceptable runtime and 89% coverage of our industrial standard cell library, including scan flip-flops, adders, and multiplexers. We also observe a significant reduction in amount of metal2 routing in comparison with industrial hand-crafted standard cells.

Standard cell libraries are an important part of many of today's integrated circuit (IC) designs. The design of all digital ASICs (Application Specific Integrated Circuit) essentially involves the use of an ASIC standard cell library comprising logic functional primitives such as basic gate functions, complex combinational functions, sequential elements, arithmetic elements and 1I0s. In this paper the work involves, designing standard cell layouts with different cells using fixed height of standard cell template and characterizing standard cells using liberate and generating. lib file for standard cell. This Standard cell library was designed an industry academia chip collaborative project. By using Cadence virtuoso, liberate tools high performance with low power consumption standard cell library was developed.

In this paper, we describe a way to build and explore auxiliary boundary constraints used to layout standard cells. It is required to build an electronic component of VLSI chip, a standard cell, given a set of constraints: geometrical design rules and additional electrical requirements. During layout of standard cells, additional constraints are formulated to avoid possible violations at cell boundaries when cells are abutted in a horizontal row. Current work proposes a flow to compute such auxiliary rules.

This paper proposes a yield optimization method for standard cells under timing constraints. Yield-aware logic synthesis and physical optimization require yield enhanced standard cells and the proposed method automatically creates yield enhanced cell layouts by decompacting the original cell layout using linear programming. We develop a novel accurate linear delay model which approximates the difference from the original delay and use this model to formulate the timing constraints into linear programming. Experimental results show that the proposed method can pick up the yield variants of a cell layout from the tradeoff curve of cell delay versus critical area and is used to create the yield enhanced cell library which is essential to realize yield-aware VLSI design flows.

In this study, we implemented an auto-correction method for layout of 45nm standard cells to get Design-for-Manufacturability (DFM) friendly design. The proposed method avoids lithography hotspot and particle defect, which are the source of systematic and random variation, by utilizing litho simulation and critical area analysis during optimization. In addition, to achieve maximum benefit of standard cell optimization, most of the recommended rules (RR) are applied to the standard cells. To assure the consistency of the optimization result, priority was given for each rule and the optimization is performed based on the priority. Using the proposed method, optimization time decreased tremendously compared to manual correction; therefore, the proposed layout optimization can handle thousands of standard cells with reasonable runtime. Verification results with manufacturing checking deck showed that DFM-friendly index improved by over 4.3% on optimized cells when compared to the original cells.

In this paper, we propose an optimal pin assignment algorithm with improvement of cell placement in standard cell layout. The objective of the algorithm is to minimize the channel density by assigning nets to terminals of cells. If the number of possible terminal assignments for each cell is bounded by some constant r, then the proposed algorithm runs in linear time. In most practical cases, the value of r is relatively small, and thus the proposed algorithm is effective and efficient to reduce the chip area.

Self-Aligned Double Patterning (SADP) is being considered for use at the 10$nm$ technology node and below for routing layers with pitches down to ~50nm because it has better LER and overlay control compared to other multiple patterning candidates. To date, most of the SADP-related literature has focused on enabling SADP-legal routing in physical design tools while few attempts have been made to address the impact SADP routing has on local, standard cell (SC) I/O pin access. In this paper, we present the first study on SADP-aware pin access and layout optimization at the SC level. Accounting for SADP-specific design rules, we propose a coherent framework that uses Mixed Integer Linear Programming (MILP) and branch and bound method to simultaneously optimize SADP-based local pin access and within-cell connections. Our experimental results show that, compared with the conventional approach, our framework effectively improves pin access of the standard cells and maximizes the pin access flexibility for routing.

In this paper, the impact of gate induced drain leakage (GIDL) on overall leakage of submicron VLSI circuits is studied. GIDL constitutes a serious constraint, with regards to off-state current, in scaled down CMOS devices for DRAM and/or EEPROM applications. Our research shows that the GIDL current is also a serious problem in scaled CMOS digital VLSI circuits. We present the experimental and simulation data of GIDL current as a function of 0.35 /spl mu/m CMOS technology parameters and layout of CMOS standard cells. The obtained results show that a poorly designed standard cell library for VLSI application may result in extremely high leakage current and poor yield.

Reliable power delivery is regarded as one of primary challenging problems in nano-meter VLSI design diagram. For dynamic voltage fluctuations on a power/ground(P/G) network, adding on-chip decoupling capacitors(decaps) is regarded as the most efficient way to reduce such noises. In this paper, we present an efficient heuristic method to allocate decaps based on transient analysis of p/g network. Our algorithm utilizes adjoint network method to efficiently calculate the sensitivity of objective function with respect to each decap value. Experimental results demonstrate that the algorithm is capable of optimizing p/g networks with hundreds of thousand nodes in 3.30 hours.

Carbon nanotube field-effect transistors (CNFETs) are promising candidates for building energy-efficient digital systems at highly-scaled technology nodes. However, carbon nanotubes (CNTs) are inherently subject to variations that reduce circuit yield, increase susceptibility to noise, and severely degrade their anticipated energy and speed benefits. Joint exploration and optimization of CNT processing options and CNFET circuit design are required to overcome this outstanding challenge. Unfortunately, existing approaches for such exploration and optimization are computationally expensive, and mostly rely on trial-and-error-based ad hoc techniques. In this paper, we present a framework that quickly evaluates the impact of CNT variations on circuit delay and noise margin, and systematically explores the large space of CNT processing options to derive optimized CNT processing and CNFET circuit design guidelines. We demonstrate that our framework: 1) runs over 100x faster than existing approaches, and 2) accurately identifies the most important CNT processing parameters, together with CNFET circuit design parameters (e.g., for CNFET sizing and standard cell layouts), to minimize the impact of CNT variations on CNFET circuit speed with less than 5% energy cost, while simultaneously meeting circuit-level noise margin and yield constraints.

Standard cells are basic building blocks, crucial for digital designs. Manufacturability improvements in standard cells have huge leverage, returning big benefits in yield and performance to all designs which may use them. Most of the benefits can be obtained by making changes to relatively few transistors in the cell, early in the design cycle. This paper presents a methodology to make standard cells more robust to manufacturing variations using DFM guidelines and knowledge of circuit performance sensitivity. Sensitivity analysis to variation parameters is performed to determine the critical cells and the critical transistors within those cells. Layout changes for DFM guidelines are performed selectively on the sensitive transistors within the cells. Results using the proposed methodology shows that there are as much as ∼2X improvement in DFM-violation score in a 45nm technology library. The proposed methodology also reduces the total-net capacitance in the cells by as much as ∼2%. These improvements are obtained without penalizing the cell area and cell pin-outs.