System and method of maximizing integrated circuit manufacturing yield with context-dependent yield cells

Abstract

A system and a method of creating context dependent yield variants of integrated circuit (“IC”) design components and using these variants during a physical design of an IC block to maximize manufacturing yield are described. A plurality of variants of each design component is generated and characterized with manufacturing yield as a function of neighboring context (“context”) that includes, but is not limited to, neighboring design components and other layout objects and shapes. The present invention describes a system and method where a physical design process, in addition to satisfying design and performance requirements such as, but not limited to, power, timing, signal integrity and minimal layout area, selects context dependent yield variants to maximize manufacturing yield.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates some embodiment of a process for assembling pre-built design components to generate an IC block.

FIG. 2 illustrates an example of logic IC block produced by an ASIC flow.

FIG. 3 illustrates some embodiment of physical design process using an ASIC flow to create an IC block.

FIG. 4 illustrates some embodiment of a memory block composed of bit cell array and peripheral blocks.

FIG. 5 illustrates some embodiment of a physical design process for generating memory block.

FIG. 6 illustrates a computing environment for the present invention.

FIG. 7 illustrates some embodiment of a process implementing the conceptual flow of the present invention.

FIG. 8 illustrates some embodiment of a process used to generate component variants for the extended component library.

FIG. 9 illustrates some embodiment of a process used to characterize a component's context dependent yield function.

FIG. 10 illustrates some embodiment of a physical design process for the present invention using context dependent yield variants to increase the manufacturing yield.

FIG. 11 illustrates an example where some embodiment of a physical design process alters neighboring context for a component.

FIG. 12 illustrates some embodiment of physical design process incorporating component context dependent yield information into an optimization objective value.

FIG. 13 illustrates some embodiment of standard cell physical design process using context dependent yield standard cell variants.

FIG. 14 illustrates some embodiment of standard cell architecture with similar layout structures near the top and bottom boundaries.

FIG. 15 illustrates some embodiment of a standard cell architecture with similar layout structures near the left and right boundaries.

FIG. 16 illustrates an example of placed standard cells with top/bottom common layout structures.

FIG. 17 illustrates the relevant left/right regions for yield characterization of standard cells with top/bottom common structures.

FIG. 18 illustrates some embodiment of process which computes the design attributes for a standard cell.

FIG. 19 illustrates some embodiment of a yield characterization process using design attribute context.

FIG. 20 illustrates some embodiment of context dependent yield standard cell variant physical design process with left/right context computation.

FIG. 21 illustrates some embodiment of context dependent yield standard cell variant physical design process with left/right design attribute computation.

FIG. 22 illustrates some embodiment of a physical design process for generating memory block using context dependent yield bit slice components.

DETAILED DESCRIPTION

In the following description, numerous details are set forth for purpose of explanation. However, one of skill in the art will realize that the invention may be practiced with variations of these specific details. In other instances, well-known structures or operations are not shown in detail to avoid obscuring the invention.

Computing Environment

FIG. 6 illustrates a block diagram of the computing environment that one embodiment of the present invention is implemented. Even though the computer system is described with specific components and architecture for illustration, it should be understood that the present invention might be implemented in several other types of embodiments. For example, the invention can be implemented on single computer with a processor chip containing 2 or more processor cores with each core containing additional hardware to maintain state of two or more threads of execution. In addition, each component can be implemented as a combination of one or more of hardware, software and firmware, even though many features of the present invention are described herein as being implemented in software.

The computing environment 600 may contain one or more components such as a communication fabric 610, random access memory (RAM) 620, central processing unit (CPU) 630, read only memory 640 (ROM), secondary memory (Storage) 650, output devices 660, input devices 670, network interface 680. All the components may communicate with each other over communication fabric 610. The communication fabric 610 collectively represents all systems, peripherals, chipset buses and all other communication pathways that can connect the components of the computing environment 600. The components of FIG. 6 are described below in further detail.

CPU 630 retrieves the instructions and data to process in order to execute the processes of this invention from the various storage components of computing environment 600. The ROM 640 stores the static instruction and data not modified during normal operation and are needed by CPU 630 and any other component of the computing environment 600. Read-write memory (RAM) 620 is a volatile storage that requires power to be supplied to store the instructions and data. Storage 650 is nonvolatile storage that doesn't need power to store instructions and data. In some embodiments, storage 650 use fixed mass-storage devices such as disk drives. Other embodiments use removal mass-storage devices such as removable disk drives. The RAM 620 stores some of the instructions and data that the CPU 630 needs. In some embodiments, the invention's processes are stored in the CPU 630, RAM 620, ROM 640, and/or storage 650.

The input device 670 enables the user to issue commands to the computing environment. Examples of an input device 670 include but are not limited to, keyboards, mouse, and/or tablet and stylus. The output device 660 is used to display images generated by the computing environment.

Network interface 680 may be implemented using protocols such as TCP/IP, ATM and/or Ethernet. In this manner, the computer can be a part of a network of computers (such as a local area network (“LAN”), a wide area network (“WAN”), or an Intranet) or a network of networks (such as the Internet). Any, some or all of the components of computing environment 600 may be used in conjunction with the invention. However, one of ordinary skill in the art would appreciate that any other system configuration may also be used in conjunction with the present invention.

As noted above, CPU 630 may retrieve the software instructions, and execute the instructions to provide various features of the present invention. The features of the present invention are described below in further detail.

General Overview

It is an object of the present invention to have a design component's manufacturing yield value be a function of the neighboring context rather than be a fixed value where a neighboring context (“context”) is defined to be neighboring design components, blocks, shapes, and other layout objects.

The present invention augments a design component library with an extended design component library with yield variants whose manufacturing yield value is context dependent rather than be a fixed value.

The present invention, in turn, allows physical design of IC blocks to utilize an extended design component library as well as a design component library to maximize a block's yield through the use of using yield variants while abiding with a IC block's design and performance goals. In some embodiment of the present invention, for IC blocks implemented by assembling pre-built design components, a plurality of variants are created and characterized for yield value as function of different neighboring context. In turn, the physical design process can query a component's yield value as function of its current context so as to allow yield to be included in a physical design optimization objectives such as, but not limited to, power, timing, and signal integrity.

In one embodiment for creating memory blocks, physical design tools are able to select an optimized set of context dependent yield variants for bit slice components when generating peripheries for a memory block.

In another embodiment for a standard cell with certain electrical characteristics, a plurality of yield variants are generated and the yield value as a function of different context is characterized. Placement and routing EDA tools can query a standard cell yield variant's expected yield value as function of surrounding context and incorporate the value when considering the other physical design optimization objectives. These EDA tools can then selectively swap out different yield variants if necessary to increase the manufacturing yield.

Conceptual Flow

FIG. 7 illustrates some embodiment of a process 7010 for a conceptual flow of the present invention. The process starts at 7030 by generating variants for each component in the Design Component Library (“DCL”) 7020 which contains the pre-constructed design components and places generated variants in a preliminary extended component library (“ECL”) 7040. At 7050, the process characterizes the yield value for each preliminary ECL as function of different context provided by components from a DCL and preliminary ECL and stores characterized yield function in a completed ECL 7060. The physical design process at 7070 accepts IC block specification 7080 along with a DCL 7020 and ECL 7060 and generates an IC block 7090 with an maximal yield value subject to other IC block requirements.

Component Variant Generation

FIG. 8 illustrates some embodiment of a process 8010 for creating variants of DCL components. The process at 8030 selects a DCL component and generates various variants. At 8040, the process generates variants of the selected DCL component. In some embodiments, variants are generated using a wide variety of methods such as, but not limited to,

• • More conservative manufacturing rules such as treating recommended rules as mandatory required rules.
  • Different layout topologies generated either using manual effort or EDA applications but keeping the electrical performance and functional behavior, the same as its corresponding DCL component.
  • Selective enforcement of some but not all recommended rules.
  • Incorporate yield learning or yield ramp data to from such sources as, but not limited to, test chips, device and/or process simulations, actual IC development or production fabrication data, and any other sources of IC manufacturing yield information.

At 8050, the process determines if there are additional DCL components from which variants haven't been generated. If there are some DCL components with no variants, the process returns to 8030 to select another DCL component. If all the DCL components have been examined, the process exits at 8060.

Context-Dependent Yield Characterization

A design component's context is the shapes within certain distance of its boundary, including shapes over within its boundary. This distance is determined by specifics of fabrication process being considered. Lithography related yield is only relevant within 5 times wavelength of the light used by the stepper. For the currently widely used 193 nm laser, this distance is around 1 um. The defect-dependent random yield impact is determined by a characteristic process particle size distribution. The yield impact dependent on mechanical mechanism such as CMP has a much larger scope, it is preferably considered during routing process. This context locality has important implications during yield characterization of design components.

FIG. 9 illustrates some embodiments for a process 9010 for characterizing context dependent yield figure of merit for DCL and ECL components. The process at 9020 selects a component, say I, from either the ECL or DCL. At 9030, the process generates a neighboring configuration composed of ECL and DCL components to completely surrounding component I so as to provide a context. At 9040, the process evaluates the manufacturing yield of component I, based on the current configuration. In some embodiments, the manufacturing yield value could be generated from several methods such as, but not limited to:

• • Printability based on lithography simulation incorporating such items, but not limited to, an aerial imaging and resist modeling across nominal and different process latitude conditions.
  • Device parameter variations due to neighboring context conditions altering localized IC fabrication process environmental conditions.
  • Inter layer dielectric (“ILD”) variability to due to chemical mechanical polishing (CMP) variability.
  • Reticle enhancement technology (“RET”) hot spots where the layout doesn't have room to accommodate an RET feature.
  • Edge placement error (“EPE”) after RET features have been inserted.

At 9050, the process determines if there are additional context configurations to be evaluated and returns to 9030 to generate another context configuration if so. If all the context configurations have been evaluated, the process stores a context dependent yield function in the ECL at 9060. At 9070, the process determines if there are additional DCL or ECL components with missing context dependent yield functions. If so, the process returns to 9020 to select another components. If all the DCL and ECL components have context dependent yield function, the process terminates at 9080.

Physical Design with Yield Variants

As illustrated in FIG. 1, prior art physical design process for generating an IC block first selects a DCL design components at 1040 and never alters a set of components through the remaining placement step at 1050 and routing step at 1060. This is due to underlying assumption that a component's yield value remaining constant during the remaining physical design process steps.

In the present invention, the premise is that a component's yield value changes as different neighboring configurations are generated and modified during a physical design process. FIG. 10 illustrates some embodiments of context dependent yield physical design process 10010 for the present invention. This flow is similar to that illustrated in FIG. 1 at 10040 where components are select from the Design Component Library 10020 based on the IC Block specification 10030. The difference is that component placement at 10050 can substitute component yield variants from the Extended Component Library 10080 whenever necessary to increase block's overall yield number during its operations. Likewise, at 10060 the Component Routing step can substitute yield variants from the Extend Component Library 10080 whenever necessary to increase a block overall yield number.

These substitutions of different context dependent yield variants occur whenever a component's context is altered during the normal operations of a physical design process and are in addition to the other physical design optimization such as, but not limited to, wire length minimization, congestion avoidance, and timing closure.

FIG. 11 illustrates an example of an intermediate placement configuration 11010 for some embodiment of a physical design process. Component 11100's context is defined by components 11020-11090 so that its yield value is computed based on the context generated by these components. Some embodiments of a physical design process produces configuration 11015 where locations originally occupied by 11020-11090 are now occupied by components 11025-11095 while component 11100, which is now labeled 11105, occupies the same location as in configuration 11010. A physical design process would first compute the “goodness” of the configuration 11015 using existing art physical design optimization objectives and then factor in the new yield value for 11105 based on its new context. In this manner, the physical design process can factor in expected manufacturing yield for a new configuration as part of its optimization objective.

FIG. 12 illustrates some embodiment of physical design optimization process 12010 to incorporate components' context dependent yield information. At 12020, the process computes a value for a “goodness” of the current physical design configuration with regards to optimization objectives using existing art approaches. At 12030, the process computes component's context dependent yield value using The system and method described in the present invention. At 12040, the process combines context yield values and existing art optimization objective value to form a composite optimization objective value. It is this new composite optimization value that allows physical design process to incorporate context dependent yield information.

Standard Cells

Some embodiments of the present invention utilize an ASIC physical design flow where design components are selected from a standard cell library. Such embodiments can take advantage of the special nature of standard cell architectures for computational efficiencies.

Standard Cell Flow

FIG. 13 illustrates a standard cell design flow process 1300 used by some embodiments of the current invention. The process maximizes the manufacture yield of integrated circuit by using context-dependent standard cell library. The flow is similar to the one shown in FIG. 3, except that instead of baseline library 310, an extended library containing context-dependent yield variants 1312 is used in the design flow in addition to a baseline library 1310.

In process 1314, logic or physical synthesis tools in modern days perform floor planning and fast placement in order to achieve timing closure, placement and routing are estimated, but they are not final. Therefore the design context information is irrelevant at this stage, process 1314 is similar to 212 in the sense that electrical characteristics such as timing, power and signal integrity is the main consideration in this process of transferring a logical design representation to a gate-level netlist of cells. A baseline library is used for process 1314, since context-dependent yield information is not irrelevant either.

Process 1316 is a detailed placement of a design, in which a placement of each cell is determined. Overlapping between cells is not allowed. In traditional flow such as one shown in FIG. 3, placement 314 is implemented, dependent on an estimated routing, to be optimized for timing, signal integrity or power. In the present invention, placement tools take into consideration of context-dependent yield in addition to these objectives, since the information about the neighbor of each cell is available when placement is completed. At 1316, placement tools replace each of baseline cells with an appropriate variant of a cell, or leave a baseline cell intact, so that the design converges to a placement solution. In the solution, the yield of each cell, as a function of its context, is consistent across the whole design, and meets the design specifications. It is preferably to make variants of same cells have same or similar footprint and electrical characteristics so that the timing, power and signal integrity specifications will not be perturbed too much, so that a convergence is easier to achieve.

Process 1318 is the detailed and final routing of a design, in which all the placed cells are connected, usually by wires on numerous conducting layers, to implement a logic function of the design. In traditional flow such as one shown in FIG. 3, routing 316 is implemented to optimize for timing, power and signal integrity. In the present invention, at 1318 routing tools take into consideration of context-dependent yield. First, conducting layers added by the routing tools should not perturb the context of placed cells if such an alternative approach exists. It adds one more check to the routing process. Second, if conducting layers added by routing tools change context of cells used in the placed design, usually when conducting layers are close to FEOL, an ECO fix is needed to swap in appropriate variants of cells to converge to a solution. The yield of each is consistent across the design, and meets the design specifications. The ECO includes an incremental placement if footprints of the replaced cells and those of their replacements are different, some rerouting if pin locations of the replaced cells and those of their replacements are different. It is preferably to make variants of same cells have same or similar footprint and electrical characteristics so that convergence is easier to be achieved.

Standard Cell Yield Characterization

For some embodiment of ASIC flow of the present invention, the context locality property discussed earlier has several implications for computational efficiency advantages for some embodiments of standard cell architectures. For these embodiments, standard cell layout styles have converged to forms where layout shapes are very similar along two or more cell boundaries. In some embodiments, top and bottom boundaries are very similar for all components in the standard cell library and are typically composed of layout structures for power and ground. FIG. 14 illustrates an example of a standard cell architecture embodiment where power and ground layout structures 14010 and 14020 are located at top and bottom boundaries. 14030-14070 are internal layout objects used to implement a standard cell. FIG. 15 illustrates example of a standard cell embodiment where the left and right cell boundaries have power and ground structures 15010 and 15020, 15030-15070 are internal layout objects used to implement a standard cell.

The next series of discussion uses as an example a standard cell architecture embodiment with similar layout shapes along the top and bottom boundaries (“top/bottom”). However, one of skill in the art will realize that similar analysis applies to standard cell embodiments with similar layout shapes along the left and right boundaries (“left/right”). The explicit description of how the analysis of top/bottom embodiment applies to that of the left/right will be omitted to avoid obscuring the descriptions.

The first implication is that top/bottom layout shapes serve as barriers so that layout shapes present on cells along the top and bottom boundaries will have minimal to no impact on a cell's yield. This implies that left and right adjacent cells are sufficient to define a standard cell's context. FIG. 16 illustrates an example of placed standard cells where cell 16050's context is defined by cell 16020 and 16080. The layout objects from cells 16010, 16030, 16040, 16070, and 16090 are shielded from those of cell 16050 by its top/bottom power and ground layout structures.

The second implication is that only the portion of the cell that is within the same distance inside a boundary is relevant for context-dependent yield. It is also recognized that shapes over the top of cell in a completed design are generally interconnects on BEOL layers, which is not relevant in the fabrication of FEOL features. Combining these facts, in the standard cells dependent design, the yield impact of design context for each cell needs only be characterized for both ends of a cell within certain distance. FIG. 17 illustrates the areas considered for cells in a standard cell dependent design. The shaded areas 1704 and 1706 should be characterized for 1702 an instance of cell A. Distance d is determined by aspects of fabrication process considered in the yield characterization. As discussed above, this distance is generally less than an I/O grid size, and in turn less than the width of any cells including filler cells. That means to calculate context-dependent yield, no further beyond the two neighboring cells has to be inspected.

To capture the characterization of yield dependent on context, a 2-dimensional table of yield number is provided for each variant of every standard cell, with the two dimensions corresponds to design contexts to the left and right ends respectively. An exemplary case to illustrate in a hypothetical format is:

cell INV1_variant1 {
yield {
table: context_table_a;
data: {
0.95, 0.97, 0.98, 0.91, 0.90, 0.98;
0.96, 0.95, 0.95, 0.93, 0.94, 0.91;
0.96, 0.94, 0.94, 0.91, 0.93, 0.92;
0.91, 0.91, 0.93, 0.94, 0.97, 0.93;
0.92, 0.94, 0.92, 0.96, 0.90, 0.98;
0.94, 0.93, 0.91, 0.95, 0.99, 0.93;
}
}
}

The context table in each dimension is dependent on the design attribute abstracted from the end zones within certain distance of each cell variants. Both distance are the same for the left and right side of a standard cell. Each dimension is exhaustive listings of all possible end zones of all cells variants in the library, including the cell it self, and an empty cell. An exemplary case is:

table {
name: context_table_a;
dimension: d_1 {
empty,
INV1_base_left,
INV1_base_right,
INV1_variant1_left,
INV1_variant1_right,
INV1_variant2_left,
INV1_variant2_right,
. . ,
INV2_base_left,
INV2_base_right,
INV2_variant1_left,
INV2_variant1_right,
INV2_variant2_left,
INV2_variant2_right,
. . ,
}
dimension: d_2 {
empty,
INV1_base_left,
INV1_base_right,
INV1_variant1_left,
INV1_variant1_right,
INV1_variant2_left,
INV1_variant2_right,
. . ,
INV2_base_left,
INV2_base_right,
INV2_variant1_left,
INV2_variant1_right,
INV2_variant2_left,
INV2_variant2_right,
. . ,
}
}

By using such a table for each cell, physical design tools are able to compute a standard cell's yield as a function of its left and right neighbors.

A further optimization is possible if one observes that a large number of standard cells share similar layout topologies at end-zone regions. For example, the input stages of cells implementing same logic functionality but with different output drive strength are usually identical; the output stages of cells with identical drive strength but different logic functionality are usually identical as well. As a result, the number of unique end zone shapes is much smaller than number of standard cell library components. This results in much smaller number of context configurations for which yield characterizations are required.

FIG. 18 illustrates the process 1800 of abstraction of design attributes of shapes in a region. A standard cell 1802 with end zone regions 1804 and 1806 provide input 1808 to the abstraction 1810. At 1810, the process generates a design attribute value contained in 1812. The possible number of values of design attributes is much smaller than the number of cells in a standard cell library. Therefore, the context table size for the cell yield is much smaller. An exemplary case for the context table is:

table {
name: context_table_a;
dimension: d_1 {
0,
1,
2,
3,
. . ,
}
dimension: d_2 {
0,
1,
2,
3,
. . ,
}
}

FIG. 19 illustrates some embodiment of a process 19010 for characterizing a standard cell component yield using design attributes. At 19020, a standard cell component is selected from the DCL and ECL. At 19030, a left and right context is generated using standard cells drawn from the DCL and ECL. At 19040 the left and right design attributes are computed. At 19050, the process determines if the design attribute contexts have undergone yield characterization. If so, the process skips to 19070. If the design attributes yield characterization is not available, the process continues to 19060 where cell's yield is characterized with context configuration. The process continues to 19070 to determine if any more context configurations have to be evaluated, and returns to 19030 if so. Otherwise, the process stores the standard cell's yield characterization in the ECL at 19080 and continues to 19090 where the process determines if there are additional ECL or DCL standard cells with missing yield characterization data. If so, the process returns to 19020. Otherwise, the process terminates at 19100.

Physical Design with Standard Cell Yield Variants

Without the use of design attributes, the physical design process using context dependent yield standard cell variants is similar to embodiment illustrated in FIG. 12. FIG. 20 illustrates some embodiments of physical design process 20010 utilizing context dependent yield variant standard cell flow. At 20020, a physical design process computes the optimization objective using existing art approaches. At 20030, the process computes left and right standard cell context and looks up the yield at 20040. At 20050, the process incorporates standard cell yield information with the existing art physical design optimization objective to form a composite optimization value.

FIG. 21 illustrates some embodiments of physical design process 21010 utilizing the context dependent yield variant flow and context design attributes. As before, the process computes an optimization objective value at 21020 using existing art approaches. At 21030, the process generates design attributes for the left and right contexts and then uses these design attributes to look up standard cell yield values at 21040. At 21050, the process combines the existing art optimization value with standard cell yield values for a composite optimization value.

Memory Blocks

Another embodiment of the present invention is the physical design flow for memory blocks where peripheries are assembled from collection of bit slice components and context dependent yield variants.

Memory Block Flow

FIG. 22 illustrates memory block physical design flow process 2210 for some embodiment of the present invention. This flow is similar to FIG. 5's process 510 except that bit slice components can be altered during the placement and routing steps. The process at 2224 accepts the memory requirements and determines the memory architecture. At 2228, the process selects a periphery to generate. At 2232, the process computes the periphery requirements and continues to 2236 where bit slice components are selected. At 2240 the process places the components and optionally swaps in yield component variants from 2254 to improve manufacturing yield. At 2244, the process routes the components and optionally swaps in yield variants from 2254 if this results in higher yield values. The process continues to 2248 where the process determines if there are more periphery to generate. If there is at least one more periphery, the process returns to 2228 to select another periphery to generate. If there are no additional peripheries, the process terminates with a completed memory block 2252.

Although the description above contains many specificities, these should be not be construed as limiting the scope of the invention but merely providing illustrations of some of the presently preferred embodiments of this invention.

Thus the scope of the invention should be determined by the appended claims and their equivalents, rather than by the examples given.

Claims

1. A method for the physical design of an integrated circuit that maximizes its yield through the use of context-dependent design component variants, comprising: receiving a collection of integrated circuit design components;generating a plurality of variants for a plurality of design components with the same electrical and functional behavior as the original design components;determining context dependent yield values for a plurality of placed design components as a function of neighboring context during the physical design of an integrated circuit; andincorporating context dependent yield values for a plurality of placed design component into physical design's cost function.

2. The method as recited in claim 1 further comprising where a plurality of placed design components can be replaced with different design component variants during physical design to improve the yield of an integrated circuit block.

3. The method as recited in claim 1 where design components are standard cells and the collection of components form a standard cell library.

4. The method as recited in claim 3 where the standard cell architecture has similar layout shapes along the top and bottom boundaries so that only left and right neighboring contexts have to be examined during context dependent yield determination.

5. The method as recited in claim 4 further comprising the steps of: generating all possible unique left and right neighboring cells for a plurality of standard cells;computing the context dependent yield value for unique combination of neighboring left and right standard cells for a plurality of standard cells;storing context dependent yield values in tables indexed with the left and right neighboring cell component identifiers for a plurality of standard cells; anddetermining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block.

6. The method as recited in claim 4 further comprising the steps of: computing design attributes for the left region of a plurality of standard cells and design attributes for the right region of a plurality of standard cells;generating all possible unique configuration of left and right neighboring standard cells for a plurality of standard cells;computing the context dependent yield values for unique combinations of neighboring left and right standard cells for a plurality of standard cells;storing context dependent yield values in a table indexed with the left and right neighboring design attribute identifiers; anddetermining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block.

7. The method as recited in claim 3 where the standard cell architecture has similar layout shapes along the left and right boundaries so that only top and bottom neighboring contexts have to be examined during context dependent yield determination.

8. The method as recited in claim 7 further comprising the steps of: generating all possible unique top and bottom neighboring cells for a plurality of standard cells;computing context dependent yield value for unique combination of top and bottom neighboring standard cells for a plurality of standard cells;storing context dependent yield values in tables indexed with the top and bottom neighboring cell component identifiers for a plurality of standard cells; anddetermining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block.

9. The method as recited in claim 7 further comprising the steps of: computing design attributes for the top region of a plurality of standard cells and design attributes for the bottom region of a plurality of standard cells;generating all possible unique top and bottom neighboring cells for a plurality of standard cells;computing the context dependent yield values for unique combinations of top and bottom standard cells for a plurality of standard cells;storing context dependent yield values in a table indexed with the top and bottom neighboring design attribute identifiers; anddetermining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block

10. The method as recited in claim 1 where the integrated design components are components composing peripheral integrated circuit blocks interfacing to memory array structures.

11. A system for the physical design of an integrated circuit that maximizes its yield through the use of context-dependent design component variants, comprising: means of receiving a collection of integrated circuit design components;means of generating a plurality of variants for a plurality of design components with the same electrical and functional behavior as the original design component;means of determining context dependent yield values for a plurality of placed design components as a function of neighboring context during the physical design of an integrated circuit; andmeans of incorporating context dependent yield data for a plurality of placed design component into physical design's cost function.

12. The system as recited in claim 11 further comprising a means where a plurality of placed design components can be replaced with different design component variants during physical design to improve the yield of an integrated circuit block.

13. The system as recited in claim 11 wherein design components are standard cells and the collection of components comprises a standard cell library.

14. The system as recited in claim 13 wherein the standard cell architecture has similar layout shapes along the top and bottom boundaries so that only left and right neighboring contexts have to be examined during context dependent yield determination.

15. The method recited in claim 14 further comprising the steps of: means of generating all possible unique left and right neighboring standard cells for a plurality of standard cells;means of computing context dependent yield values for unique combination of neighboring left and right standard cells for a plurality of standard cells;means of storing context dependent yield values in tables indexed with the left and right neighboring cell component identifiers for a plurality of standard cells; andmeans of determining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block.

16. The system as recited in claim 14 further comprising the steps of: means of computing design attributes for left regions of a plurality of standard cells and design attributes for right regions of a plurality of standard cells;means of generating all possible unique configuration of left and right neighboring standard cells for a plurality of standard cells;means of computing the context dependent yield values for unique combinations of neighboring left and right standard cells for a plurality of standard cells;means of storing context dependent yield values in a table indexed with the left and right neighboring design attribute identifiers; andmeans of determining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block.

17. The system as recited in claim 13 wherein the standard cell architecture has similar layout shapes along the left and right boundaries so that only top and bottom neighboring contexts have to be examined during context dependent yield determination.

18. The system as recited in claim 17 further comprising the steps of: means of generating all possible unique top and bottom neighboring standard cells for a plurality of standard cells;means of computing context dependent yield values for unique combination of neighboring top and bottom standard cells for a plurality of standard cells;means of storing context dependent yield values in tables indexed with the top and bottom neighboring cell component identifiers for a plurality of standard cells; andmeans of determining components' context dependent yields by performing lookups of these tables during physical design of an integrated circuit block.

19. The system as recited in claim 17 further comprising the steps of: means of computing design attributes for top regions of a plurality of standard cells and design attributes for bottom regions of a plurality of standard cells;means of generating all possible unique configuration of neighboring top and bottom standard cells for a plurality of standard cells;means of computing the context dependent yield values for unique combinations of neighboring top and bottom standard cells for a plurality of standard cells;means of storing context dependent yield values in a table indexed with the top and bottom neighboring design attribute identifiers; andmeans of determining components' context dependent yields by performing lookups of these 1 tables during physical design of an integrated circuit block

20. The system as recited in claim 11 where the integrated design components are components composing peripheral integrated circuit blocks interfacing to memory array structures.