Method and apparatus for representing items in a design layout

FIELD OF THE INVENTION

[0002] The invention is directed towards method and apparatus for representing items in a design layout.

BACKGROUND OF THE INVENTION

[0003] Design engineers design IC's by transforming circuit description of the IC's into geometric descriptions, called layouts. To create layouts, design engineers typically use electronic design automation (“EDA”) applications. These applications provide sets of computer-based tools for creating, editing, and analyzing IC design layouts. EDA applications create layouts by using geometric shapes that represent different materials and devices on IC's. For instance, EDA tools commonly represent IC components as rectangular objects. They also use rectangles to represent horizontal and vertical interconnections between rectangular IC components. The horizontal and vertical interconnects are referred to as Manhattan interconnects or wiring.

[0004] EDA tools typically represent a rectangular object in terms of x and y coordinates of two opposing vertices of the object (i.e., in terms of xLO, yLO, and xHI, yHI). These tools often represent a non-rectangular object in terms of the x- and y-coordinates of the object's vertices. Such an approach works well when most of the shapes in the layout are rectangular.

[0005] However, certain EDA tools have recently been proposed that support non-Manhattan wiring. These tools often utilize a large number of non-rectangular shapes. The traditional approach of specifying a shape in terms of the coordinates of two or more of its vertices is not efficient for these newer EDA tools. Therefore, there is a need for a novel way of specifying items in a design layout that has non-rectangular shapes.

[0006] There is also a need for an efficient way of representing routes with non-Manhattan edges. One prior technique for representing a Manhattan route specifies the route in terms of a set of segments and one or more styles. Each segment is a straight line that connects two points in the design layout. In some cases, the segments of a route are run-length encoded. This encoding specifies a direction and a length for each segment. In such prior encoding, a segment's direction can be along 0°, 90°, 180°, and 270°. This encoding also specifies an order for the segments of the route.

[0007] Each segment's style specifies four values that can be used to transform a line-representation of the segment into a rectangular shape that is a more complete geometric representation of the segment. These four values include two low delta values, dxLO and dyLO, and two high delta values, dxHI, and dyHI. The low delta values are subtracted from the segment's lowest x- and y-values xLO, yLO to obtain the low x- and y-values (xRecLO, yRecLO) of rectangle that represents the segment, while the high delta values are added to the segment's highest x- and y-values xHI, yHI to obtain the high x- and y-values (XRecHl, YRecHI) of the rectangle. FIG. 1 illustrates an example of constructing the rectangle for a segment from the coordinates and style values of the segment. This prior technique, however, does not support non-Manhattan routes. Accordingly, there is a need for an efficient representation of routes that can have non-Manhattan edges.

SUMMARY OF THE INVENTION

[0008] Some embodiments of the invention provide novel methods for representing items in a design layout. For instance, some embodiments use a method that represents an item in terms of n values that define n half-planes, which when intersected define the shape of the item. In some embodiments, n is a number greater than 4.

[0009] Some embodiments use a method that (1) identifies a first set of location data for the item with respect to a first coordinate system, (2) identifies a second set of location data for the item with respect to a second coordinate system, and (3) specifies the item in terms of the first and second set of location data. In some embodiments, both the first and second coordinate systems have first and second coordinate axes. In some of these embodiments, the first set of location data are the highest and lowest coordinates of the item on the first and second coordinate axes of the first coordinate system, while the second set of location data are the highest and lowest coordinates of the item on the first and second coordinate axes of the second coordinate system.

[0010] Some embodiments use a method that (1) receives a first set of data that defines the item with respect to a first coordinate system of the design layout, (2) from the first set of data, generates a second set of data that defines the item with respect to a second coordinate system of the design layout, and (3) specifies the item in terms of both the first and second sets of data.

[0011] In some of the embodiments mentioned above, the first coordinate system is a Manhattan coordinate system, while the second coordinate system is a non-Manhattan coordinate system. In other embodiments, the first coordinate system is a non-Manhattan coordinate system, while the second coordinate system is a Manhattan coordinate system.

BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The novel features of the invention are set forth in the appended claims. However, for purpose of explanation, several embodiments of the invention are set forth in the following figures.

[0013]
FIG. 1 illustrates an example of constructing the rectangle for a segment from the coordinates and style values of the segment.

[0014]
FIG. 2 illustrates an octangle data structure used in some embodiments of the invention.

[0015]
FIG. 3 illustrates two coordinates systems used to define the parameters of the octangle data structure.

[0016]
FIG. 4 illustrates a −45° diagonal line that can be represented by an equation that is defined by reference to one of the coordinate systems illustrated in FIG. 3.

[0017]
FIGS. 5-12 illustrate eight half planes that are defined by the eight parameters of the octangle data structure illustrated in FIG. 2.

[0018]
FIG. 13 illustrates examples of the types of shapes that the octangle data structure of FIG. 2 can represent.

[0019]
FIGS. 14-22 illustrate how the shapes illustrated in FIG. 13 are represented by the eight half-plane values.

[0020]
FIGS. 23A and 23B illustrate an example that illustrated why some embodiments define a superfluous half plane of the octangle data structure to abut one of the vertices of the geometric shape defined by the octangle data structure.

[0021]
FIG. 24 illustrates a process that defines an octangle data structure for a design-layout item that is initially defined in terms of the x- and y-coordinates of its vertices.

[0022]
FIG. 25 illustrates a process that defines an octangle data structure for a design-layout item that is initially defined in terms of the s- and t-coordinates of its vertices.

[0023]
FIGS. 26 and 27 illustrate the use of two different styles for the same segment.

[0024]
FIG. 28 illustrates a process for constructing the polygonal shape of a wire segment from its direction, length, and associated style information.

[0025]
FIG. 29 illustrates a data structure and database arrangement that is used in some embodiments in conjunction with the process of FIG. 28.

[0026]
FIG. 30 illustrates two different view of a design layout.

[0027]
FIGS. 31 and 32 illustrate a process that generates the segment and style information from a polygon that represents a route segment.

[0028]
FIG. 33 illustrates an example of identifying the intersection of a triangle and an octagon.

[0029]
FIG. 34 illustrates an example of identifying a union of a rectangle and an octagon.

[0030]
FIG. 35 illustrates a portion of a polygon that is being reduced.

[0031]
FIG. 36 illustrates an example of constructing an octilinear bounding polygon for three design-layout items, which are a rectangle, a triangle, and a square.

[0032]
FIG. 37 provides an example of the intersection of two lines that represent two edges of two half plane.

[0033]
FIG. 38 conceptually illustrates a computer system with which some embodiments of the invention are implemented.

[0034]
FIGS. 39-43 illustrate some alternative data structures and coordinate systems that are used by some embodiments of the invention.

DETAILED DESCRIPTION OF THE INVENTION

[0035] In the following description, numerous details are set forth for purpose of explanation. However, one of ordinary skill in the art will realize that the invention may be practiced without the use of these specific details. In other instances, well-known structures and devices are shown in block diagram form in order not to obscure the description of the invention with unnecessary detail.

[0036] Some embodiments of the invention provide novel methods for representing items in a design layout. For instance, some embodiments use a method that represents an item in terms of n-values that define n-half planes, which when intersected define the shape of the item. In some embodiments, n is a number greater than four.

[0037] Some embodiments use a method that (1) identifies a first set of location data for the item with respect to a first coordinate system, (2) identifies a second set of location data for the item with respect to a second coordinate system, and (3) specifies the item in terms of the first and second set of location data. In some embodiments, both the first and second coordinate systems have first and second coordinate axes. In some of these embodiments, the first set of location data are the highest and lowest coordinates of the item on the first and second coordinate axes of the first coordinate system, while the second set of location data are the highest and lowest coordinates of the item on the first and second coordinate axes of the second coordinate system.

[0038] Some embodiments use a method that (1) receives a first set of data that defines the item with respect to a first coordinate system of the design layout, (2) from the first set of data, generates a second set of data that defines the item with respect to a second coordinate system of the design layout, and (3) specifies the item in terms of both the first and second sets of data.

[0039] In some of the embodiments mentioned above, the first coordinate system is a Manhattan coordinate system, while the second coordinate system is a non-Manhattan coordinate system. In other embodiments, the first coordinate system is a non-Manhattan coordinate system, while the second coordinate system is a Manhattan coordinate system.

[0040] I. Octangle Data Structure

[0041] Some embodiments of the invention provide novel data structures and processes that efficiently store convex geometric shapes in a design layout, when at least some convex geometric shapes have at least two sides that are neither parallel nor orthogonal to each other. In some embodiments, each possible side of the shapes is one of the manufacturing directions. As used in this document, the term “manufacturing” refers to the process used to manufacture an IC from its EDA design. Hence, manufacturing directions are the directions available to the manufacturing process to produce an IC from an EDA design (e.g., to specify the shapes of cells, pins, interconnects, etc. from the EDA design).

[0042]
FIGS. 2-25 illustrate the data structure and processes of some embodiments. The data structure illustrated in these figures is referred to below as the octangle data structure, as it is useful for design layouts that have items with horizontal, vertical, and/or ±45° directions. In the description below, both the wiring and non-wiring geometries of the design layout are convex shapes, or can be decomposed into convex shapes, that can have horizontal, vertical, and ±45° sides. One of ordinary skill will realize, however, that some embodiments might use the octangle data structure in cases where the wiring or non-wiring geometries are more restricted. For instance, this data structure might be used when the wiring is only in the Manhattan directions or in the ±45° directions, or the non-wiring geometries (e.g., pins and obstacles) have sides along only in the Manhattan directions or the ±45° directions. The discussion below focuses on convex polygonal shapes as it assumes that any non-convex shapes can be broken down into convex polygonal shapes.

[0043] As illustrated in FIG. 2, the octangle data structure 205 represents each convex geometric shape in terms of eight values, xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI. These eight values define eight half plane in two coordinate systems. Before describing how 8-half planes define a convex geometric shape, the two coordinate systems are briefly described. FIG. 3 illustrates the two coordinates systems. As shown in this figure, one coordinate system is a Manhattan coordinate system 305, which is formed by an x-axis 310 and y-axis 315. The other coordinate system is a 45′-rotated coordinate system 320, which is formed by an s-axis 325 and t-axis 330. The s-axis is at a 45° counterclockwise rotation from the x-axis 310, while the t-axis is at a 135° counterclockwise rotation from the x-axis 310. In the layouts of some embodiments, horizontal lines are aligned with the x-axis, vertical lines are aligned with the y-axis, 45° diagonal lines are aligned with the s-axis, and −45° diagonal lines are aligned with the t-axis.

[0044] In some embodiments, the Manhattan coordinate system is aligned with the manufacturing grid. Also, some embodiments define the unit length along the s- and t-axes to be

\frac{\sqrt{2}}{2}

[0045] times the unit length along the x- and y-axes, so that the diagonal coordinate system is also aligned with manufacturing grid. In other words, the coordinate resolution along the diagonal coordinate system 320 is selected such that diagonal lines that traverse through integer diagonal coordinates can start and terminate on the manufacturing grid, and can intersect, on the manufacturing grid, the Manhattan lines that traverse through integer Manhattan coordinates. For instance, FIG. 4 illustrates a −45° diagonal line 405 that can be represented by the equation,

s=5.

[0046] This line 405 intersects the Manhattan coordinate system at (0, 5) and (5,0). As shown in FIG. 4, 5 units along the s-axis is equivalent to

5 * \frac{\sqrt{2}}{2},

[0047] or 3.54, units along the Manhattan coordinate system. Given that the unit length along the s- and t-axes is

\frac{\sqrt{2}}{2}

[0048] times the unit length along the x- and y-axes, the following coordinate-transform equations (A) and (B) can be used to obtain the s- and t-coordinates of any point from the points x- and y-coordinates.

s=x+y
, and (A)

t=y−x.
(B)

[0049] As mentioned above, the octangle data structure 205 represents each convex geometric shape in terms of 8 values, where each value defines a half plane that is defined in the design layout by reference to one of the two coordinate systems 305 and 320. FIG. 2 identifies these 8 values as xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI. The values xLO, yLO, sLO, tLO are the smallest coordinates of the shape along the x-, y-, s-, and t-axes (i.e., the smallest x-, y-, s-, and t-axis values of the shape), while xHI, yHI, sHI, and tHI are the largest coordinates of the shape along the x-, y-, s-, and t-axes (i.e., the largest x-, y-, s-, and t-axis values of the shape). As illustrated in FIGS. 5-12, the low values (i.e., xLO, yLO, sLO, tLO) specify a half-plane to the positive side (i.e., the right or top side) of the value, while the positive values (i.e., xHI, yHI, sHI, and tHI) specify a half-plane to the negative side (i.e., the left or bottom side) of the value. (Each of these figures illustrates its half plane as a set of dashed lines that terminate on a solid line.) The intersection of these 8 half-planes defines a convex geometric shape. Specifically, the eight values xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI define a convex geometric shape

\underset{\in ℛ^{2}}{S (x, y)}

[0050] where:

xLO≦x≦xHI,

yLO≦y≦yHI,

sLO≦s≦sHI,

tLO≦t≦tHI,

s=x+y, and

t=y−x.

[0051]
FIG. 13 illustrates examples of the types of shapes that the octangle data structure 205 can represent. As shown in this figure, this data structure can represent a single point, a line, or any convex polygon that has anywhere from three to eight sides in 8 possible directions. FIGS. 14-22 illustrate how the shapes illustrated in FIG. 13 are represented by the 8 half-plane values xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI. (To simplify these drawings, FIGS. 14-22 illustrate each half plane by a dashed line that represents the line on which the half plane terminates.) The polygonal shapes illustrated in FIGS. 13-22 can represent any number of items in the design layout. These items include cells, circuits, circuit components (such as devices like transistors, capacitors, etc.), obstacles, vias, pins, bounding polygons, regions of interest, routes, etc. For instance, the polygon in FIG. 16 might be all or a portion of a cell, a circuit, a circuit component, an obstacle, a pin, a via, a bounding polygon, a region of interest, or a route, etc.

[0052] When a convex geometric shape has less than eight sides, one or more of the eight half planes of the octangle data structure are superfluous (i.e., are not necessary to uniquely identify the boundaries of the shape). For instance, the half plane tHI is superfluous in FIG. 15, the half planes tHI and sLO are superfluous in FIG. 16, the half planes tHI, tLO, and sLO are superfluous in FIG. 17, and so on. When one of the eight half planes is superfluous, it is canonically defined to abut one of the vertices of the geometric shape. This canonical definition is part of the above-mentioned definition of the eight half-plane values xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI. As mentioned above, the values xLO, yLO, sLO, tLO are defined as the smallest coordinates of the shape along the x-, y-, s-, and t-axes (i.e., the smallest x-, y-, s-, and t-axis values of the shape), while xHI, yHI, sHI, and tHI are defined as the largest coordinates of the shape along the x-, y-, s-, and t-axes (i.e., the largest x-, y-, s-, and t-axis values of the shape).

[0053] Some embodiments define a superfluous half plane to abut one of the vertices of the geometric shape in order to ensure that future calculations that are performed on the shape provide accurate results. FIGS. 23A and 23B illustrate an example of this. These figures present the vertex 2305 of a geometric shape 2310 and a superfluous half plane 2315 that is defined by the sHI parameter. In FIG. 23A, the sHI for the half plane 2315 correctly abuts the vertex 2305. Hence, as shown in FIG. 23A, an expansion operation that expands the sides of the geometric shape 2310 by a particular factor, results in a new geometric shape 2320 that correctly has a −45° side (due to the expansion of the half plane 2315) on its upper right hand corner. FIG. 23B, on the other hand, illustrates the sHI for the half plane 2315 to be incorrectly positioned away from the vertex 2305. Consequently, an expansion operation on this representation incorrectly results in a new geometric shape 2325 that still has a 90° juncture at the upper left corner of the new geometric shape 2325. Expansion operations will be further described below.

[0054] Another advantage of defining a superfluous half plane to abut one of the vertices of the geometric shape is that this definition simplifies the identification of bounding polygons and the computation of their widths. As further described below, the octangle data structure allows EDA tools to quickly identify the octilinear bounding polygon for a set of items in the design layout. From the octilinear bounding polygon, the rectilinear Manhattan bounding box can be quickly identified by simply taking the low and high x and y-values xLO, yLO, xHI, yHI of the octilinear bounding polygon. Similarly, the 45′-rotated bounding box can be quickly identified by simply taking the low and high s- and t-values sLO, tLO sHI, and tHI of the octilinear bounding polygon. Also, the width of each of these bounding polygons in a particular direction can be quickly computed by subtracting the box's low value from the box's high value in the particular direction.

[0055] The octangle data structure 205 can be used to define design-layout items that are in shape of points, lines, or a variety of convex polygons. Examples of such items include cells, circuits, circuit components (such as devices like transistors, capacitors, etc.), vias, obstacles, pins, bounding polygons, regions of interest, and routes. When a design-layout item is represented by a non-convex shape that is decomposable into two or more convex shapes, the design-layout item is defined by two or more octangle data structures, where each octangle data structure represents one of the convex shapes that forms the item.

[0056] Some embodiments use essentially similar processes to create the octangle data structure for cells, circuits, circuit components, vias, obstacles, pins, bounding polygons, and regions of interest, but use a slightly more efficient approach for representing routes. In the description below, the creation of the more general octangle data structure will be first described below. This description will then be followed by a description of the efficient representation of routes in some embodiments of the invention.

[0057]
FIG. 24 illustrates a process 2400 that defines an octangle data structure for a design-layout item that is initially defined in terms of the x- and y-coordinates of its vertices. The design-layout item can be any item in the layout, such as a cell, a circuit, a circuit component, a via, an obstacle, a pin, a bounding polygon, a route and a region of interest. In some cases, the item is defined only in terms of two of its vertices. For instance, in some cases, the item is a rectangle that is initially defined only in terms of x- and y-coordinates of two of its opposing corners (i.e., only in terms of xLO, yLO, and xHI, yHI). Alternatively, the item might be defined in terms of three or more of its vertices. For example, the item might be a polygon that is defined in terms of the x- and y-coordinates of each of its vertices.

[0058] As shown in FIG. 24, the process 2400 initially defines (at 2405) an octangle data structure 205 for the design-layout item. At 2410 and 2415, the process then identifies the lowest x- and y-values of the item, and stores these values as the xLO and yLO values in the octangle data structure 205. The item's lowest x- and y-values are respectively the smallest x-coordinate and the smallest y-coordinate of the item's vertices.

[0059] At 2420 and 2425, the process then identifies the highest x- and y-values of the item, and stores these values as the xHI and yHI values in the octangle data structure 205. The item's highest x- and y-values are respectively the largest x-coordinate and the largest y-coordinate of the item's vertices.

[0060] The process then performs (at 2430) a coordinate-transform operation. From the x- and y-coordinates of the vertices of the design-layout item, the coordinate-transform operation (at 2430) produces an s- ant t-coordinates for each vertex of the design-layout item. The process produces the s- and t-coordinates for a particular vertex from the x- and y-coordinates of the particular vertex by using the above-described coordinate-transform equations (A) and (B). Also, when the item was initially defined in terms of the x- and y-coordinates of only two of its vertices, the process initially identifies (at 2430) the x- and y-coordinates of the other vertices of the item, and then performs (at 2430) its coordinate-transform operation.

[0061] At 2435 and 2440, the process then identifies the smallest s- and t-values of the vertices of the item (i.e., the smallest s- and t-values computed at 2430), and stores these values as the sLO and tLO values in the octangle data structure 205. At 2445 and 2450, the process next identifies the largest s- and t-values of the vertices of the item (i.e., the largest s- and t-values computed at 2430), and stores these values as the sHI and tHI values in the octangle data structure 205. After 2450, the process ends.

[0062] The process 2400 of FIG. 24 can also be used to define an octangle data structure that represents an octilinear-bounding polygon of a set of design-layout items that are each defined in terms of the x- and y-coordinates of two or more vertices. When the process 2400 is used in such a manner, the process 2400 in some embodiments (1) identifies the bounding polygon's xLO, yLO, xHI and yHI values by examining (at 2410-2425) the vertices of all the items in the set of design-layout items, (2) performs (at 2430) coordinate transform operations on all the vertices of all the items in the set, and (3) identifies the bounding polygon's sLO, tLO, sHI and tHI values by examining (at 2435-2450) the vertices of all the items in the set of design-layout items.

[0063]
FIG. 25 illustrates a process 2500 that defines an octangle data structure for a design-layout item that is initially defined in terms of the s- and t-coordinates of its vertices. The design-layout item can be any item in the layout, such as a cell, a circuit, a circuit component, a via, an obstacle, a pin, a bounding polygon, a route, a region of interest, etc. In some cases, the item is defined only in terms of two of its vertices. For instance, in some cases, the item is a rotated rectangle that is initially defined only in terms of s- and t-coordinates of two of its opposing corners (i.e., only in terms of sLO, tLO, and sHI, tHI). Alternatively, the item might be defined in terms of three or more of its vertices. For example, the item might be a polygon that is defined in terms of the s- and t-coordinates of each of its vertices.

[0064] As shown in FIG. 25, the process 2500 initially defines (at 2505) an octangle data structure 205 for the design-layout item. At 2510 and 2515, the process then identifies the lowest s- and t-values of the item, and stores these values as the sLO and tLO values in the octangle data structure 205. The item's lowest s- and t-values are respectively the smallest s-coordinate and the smallest t-coordinate of the item's vertices.

[0065] At 2520 and 2525, the process then identifies the highest s- and t-values of the item, and stores these values as the sHI and tHI values in the octangle data structure 205. The item's highest s- and t-values are respectively the largest s-coordinate and the largest t-coordinate of the item's vertices.

[0066] The process then performs (at 2530) a coordinate-transform operation. From the s- and t-coordinates of the vertices of the design-layout item, the coordinate-transform operation (at 2530) produces an x- ant y-coordinates for each vertex of the design-layout item. The process produces the x- and y-coordinates for a particular vertex from the s- and t-coordinates of the particular vertex by using the following coordinate-transform equations (C) and (D), which are derived from the above-described equations (A) and (B).

x=
½(s−t), and (C)

y=
½(s+t). (D)

[0067] Also, if the item was initially defined in terms of the s- and t-coordinates of only two of its vertices, the process initially identifies (at 2530) the s- and t-coordinates of the other vertices of the item, and then performs (at 2530) its coordinate-transform operation.

[0068] At 2535 and 2540, the process then identifies the smallest x- and y-values of the vertices of the item (i.e., the smallest x- and y-values computed at 2530), and stores these values as the xLO and yLO values in the octangle data structure 205. At 2545 and 2550, the process next identifies the largest x- and y-values of the vertices of the item (i.e., the largest x- and y-values computed at 2530), and stores these values as the xHI and yHI values in the octangle data structure 205. After 2550, the process ends.

[0069] The process 2500 of FIG. 25 can be used to define an octangle data structure that represents an octilinear-bounding polygon of a set of design-layout items that are each defined in terms of the s- and t-coordinates of two or more vertices. When the process 2500 is used in such a manner, the process 2500 in some embodiments (1) identifies the polygon's sLO, tLO, sHI and tHI values by examining (at 2510-2525) the vertices of all the items in the set of design-layout items, (2) performs (at 2530) coordinate transform operations on all the vertices of all the items in the set, and (3) identifies the polygon's xLO, yLO, xHI and yHI values by examining (at 2535-2550) the vertices of all the items in the set of design-layout items.

[0070] It might be necessary to identify the octilinear-bounding polygon of a set of design-layout items that includes (1) a first sub-set of items that are defined in terms of the x- and y-coordinates of two or more of their vertices, and (2) a second sub-set of items that are defined in terms of the s- and t-coordinates of two or more of their vertices. In such a case, some embodiments perform two sets of coordinate transform operations initially to obtain x-, y-, s- and t-vertex coordinates for each item in the set. These embodiments then identify the smallest and largest x-, y-, s- and t-vertex coordinate values in the set and store these values in an octangle data structure. The initial coordinate-transform operations are not needed when each item in the set is expressed in terms of x-, y-, s-, and t-coordinates.

[0071] In some cases, some embodiments do not perform the process 2400 or the process 2500 to identify the octangle data structure of a design-layout item. Instead, in such cases, these embodiments define the octangle data structure while defining the design-layout item. For instance, such is the case when some embodiments define an item based on a user's input.

[0072] II. Representation of Routes

[0073] Some embodiments of the invention define each route in terms of one or more segments and one style for each segment. Each segment is a straight line that connects two points in the design layout. Some embodiments run-length encode the segments of a route. Such an encoding specifies each segment in terms of a direction D and a length L. A segment's direction can be along any of the 8 available routing directions, which are 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315°. This encoding also specifies an order for the segments of the route. Such an order can be specified in a variety of ways, such as a singly or doubly linked list, a numbering scheme, etc. In some embodiments, the segments are stored in a 1-dimensional array. In these embodiments, the order of the segments is the order that they appear in the array.

[0074] The run-length encoding of a route also typically specifies at least a starting point for the route. From this starting point, a line representation of the route can be constructed as follows. The first segment S1 is initially constructed as a segment that starts at the starting point and terminates a distance d away from this starting point in the first segment's direction D1. When the first segment is in a horizontal or vertical direction, the distance d equals the length L1 of the first segment. On the other hand, when the first segment is in a ±45° direction, the distance d equals the length L1 of the first segment times square root of 2 (i.e., d=L1*{square root}{square root over (2)}).

[0075] Each subsequent segment of the route is then constructed by performing an iterative operation. During each iteration, the operation constructs one segment, referred to as the current segment SC. This current segment SC starts at the termination of the previously constructed segment SP, which is before the current segment in the route in the segment order that is being constructed. The current segment SC terminates at a point that is a distance dC (where dC is the current segment's length LC when the segment is in a horizontal or vertical direction, or is the current segment's length LC*{square root}{square root over (2)}when the segment is in a ±45° direction) away from the previous segment's termination point in the current segment's direction D. This iterative operation terminates once it has constructed the last segment.

[0076] As mentioned above, some embodiments also define a style for each segment of a route. A style specifies eight values that can be used to transform a line-representation of a segment into a convex polygonal shape that represents a more complete geometric representation of the segment. These eight values include four low values, dxLO, dyLO, dsLO, dtLO, and four high values, dxHI, dyHI, dsHI, and dtHI. The low values are subtracted from the segment's lowest x-, y-, s- and t-values to obtain the low values of convex polygon that represents the segment, while the high values are added to the segment's highest x-, y-, s-, and t-values to obtain the high values of the convex polygon.

[0077]
FIGS. 26 and 27 illustrate the use of two different styles for the same segment 2600. The segment 2600 connects the x,y coordinates (1,3) and (5,7), which map to the s,t coordinates (4,2) and (12,2). In the example illustrated in FIG. 26, the style is specified by the following eight values:

dxLO=2, dyLO=2, dsLO=3, dtLO=3, dxHI=2, dyHI=2, dsHI=3, dtHI=3.

[0078]
FIG. 26 illustrates that the subtraction of the style's low values from the corresponding low values of the segment 2600 identifies the following four half-planes:

xLO=−1, yLO=1, sLO=1, tLO=−1,

[0079] while the addition of the style's high values from the corresponding high values of the segment 2600 identifies the following four half-planes:

xHI=7, yHI=9, sHI=15, tHI=5.

[0080] As shown in FIG. 26, these eight half planes define an octagon 2605.

[0081] In the example illustrated in FIG. 27, the style is specified by the following eight values:

dxLO=2, dyLO=2, dsLO=3, dtLO=3, dxHI=2, dyHI=2, dsHI=1, dtHI=3.

[0082]
FIG. 27 illustrates that the subtraction of the style's low values from the corresponding low values of the segment 2600 identifies the following four half-planes:

xLO=−1, yLO=1, sLO=1, tLO=−1,

[0083] while the addition of the style's high values from the corresponding high values of the segment 2600 identifies the following four half-planes:

xHI=7, yHI=9, sHI=13, tHI=5.

[0084] As shown in FIG. 27, these eight half planes define a hexagon 2705. Accordingly, in the example illustrated in FIGS. 26 and 27, the use of two different styles for the same segment 2600 resulted in two different convex polygonal shapes 2605 and 2705 for the segment. FIG. 28 illustrates a process for constructing the polygonal shape of a wire segment from its direction, length, and associated style information. This process works in conjunction with a data structure and database arrangement illustrated in FIG. 29. FIG. 29 illustrates that a segment is represented as a segment data structure 2905 that stores the length and direction of the segment, and potentially other segment identifying indicia, such as relational data that specifies the order of the segment in the segment's route. As shown in this figure, the segment data structure also includes an index 2910 that identifies a record 2920 in a style table 2915. Each record in the style table 2915 is identifiable by a unique style index 2910. Also, each record specifies a unique set of 8 delta values, dxLO, dyLO, dsLO, dtLO, dxHI, dyHI, dsHI, and dtHI.

[0085] As shown in FIG. 28, the process 2800 initially identifies (at 2805) the x-, y-, s-, and t-coordinates of the start and end points of the segment from the segment's start coordinate, direction, and length. As mentioned above, a particular segment's start coordinate is the termination coordinate of the previous segment in the route, unless the particular segment is the first segment, in which case its start coordinate is the start point of the route. The segment's end point is a point that is a distance d away from the segment's start point in the segment's direction D. As mentioned above, the distance d is the segment's length L when the segment is in a horizontal or vertical direction, or is the segment's length times square root of 2(L*{square root}{square root over (2)}) when the segment is in a ±45° direction.

[0086] The process then retrieves (at 2810) the delta values (dxLO, dyLO, dsLO, dtLO, dxHI, dyHI, dsHI, and dtHI) associated with the style of the segment. It then computes (at 2815) the x-low value xPLO of the polygon that is to represent the segment by subtracting dxLO from the segment's xLO. The process next computes (at 2820) the y-low value yPLO of the polygon by subtracting dyLO from the segment's yLO. It then computes (at 2825) the s-low value sPLO of the polygon by subtracting dsLO from the segment's sLO. The process next computes (at 2830) the t-low value tPLO of the polygon by subtracting dtLO from the segment's tLo.

[0087] The process then computes (at 2835) the x-high value xPHI of the polygon that is to represent the segment by adding dxHI to the segment's xHI. The process next computes (at 2840) the y-high value yPHI of the polygon by adding dyHI to the segment's yHI. It then computes (at 2845) the s-high value sPHI of the polygon by adding dsHI to the segment's sHI. The process next computes (at 2850) the t-high value tPHI of the polygon by adding dtHI to the segment's tHI. The computed parameters xPLO, yPLO, sPLO, tPLO, xPHI, yPHI, sPHI, and tPHI for the polygon define the shape of the polygon. Accordingly, these values can be stored in an octangle data structure to represent the polygon.

[0088] The process 2800 can be used in a variety of different scenarios when it is desirable to specify the actual geometry of the routes from their segment and style information. For instance, this process could be used to generate a display of the routes, to output data to a GDS file, to perform extraction, to estimate routing capacity, etc.

[0089]
FIG. 30 illustrates two different views of a design layout 3000 that includes three routes 3005, 3010, and 3015. The first view 3020 illustrates each route in terms of a set of line segments that are abutting. The second view 3025 illustrates each route in terms of a set of polygons that partially overlap. Each of the polygons in the second view corresponds to a particular segment in the first view. As described above by reference to FIG. 28, each particular polygon in the second view can be generated from its corresponding segment by applying the delta parameters of the segment's associated style to the style's endpoint values. Hence, to store efficiently the routes illustrated in FIG. 30, some embodiments (1) store each route in terms of the route's set of segments, and (2) as illustrated in FIG. 29, specify each segment in terms of its length, direction, and style-table index. The style table index of a segment, in turn, identifies a set of delta parameters, which when applied to the segment's endpoints specify eight half planes that define the segment's polygonal shape.

[0090]
FIG. 31 illustrates a process 3100 that generates the segment and style information from a polygon that represents a route segment. An EDA tool performs this process when the tool does not have the style and segment information. For instance, a power router might perform such an operation if it does not use a pre-defined style for the routes that it defines. It should be noted that when an EDA tool has the style and/or segment information associated with a polygonal representation of a route, the tool simply generates the needed segment and style information from the eight half planes that represent the polygon and the available segment and style information.

[0091] As shown in FIG. 31, the process 3100 initially identifies (at 3105) the coordinate axis that will be parallel to the segment that the process will define for the polygon that represents the route segment. FIG. 32 illustrates how some embodiments identify this coordinate axis. This figure illustrates a polygon 3205 for which segment and style data needs to be generated. For the polygon 3205, some embodiments identify the Manhattan bounding box 3210 and 45′-rotated bounding box 3215 that enclose the polygon. These bounding boxes can be identified according to the approach described below.

[0092] These embodiments next identify the bounding box with the smallest area, and then based on the identified bounding box, specify the coordinate axis along which the route segment will be defined. For instance, when the Manhattan bounding box has the smaller area, then the process 3100 determines (at 3105) whether the polygon's width is longer along the x-axis than its width along the y-axis. If so, the process identifies (at 3105) the x-axis as the coordinate axis. Otherwise, the process identifies (at 3105) the y-axis as the coordinate axis. On the other hand, when the 45′-rotated bounding box has the smaller area, then the process determines (at 3105) whether the polygon's width is longer along the s-axis than its width along the t-axis. If so, the process identifies (at 3105) the s-axis as the coordinate axis. Otherwise, the process identifies (at 3105) the t-axis as the coordinate axis. In the example illustrated in FIG. 32, the rotated bounding box 3215 has the smaller area and the polygon's width along the s-axis is larger than its width along the t-axis. Accordingly, the s-axis is identified (at 3105) as the coordinate axis for the segment that is to be defined for the polygon 3205.

[0093] After identifying the coordinate axis at 3105, the process defines (at 3110) the segment as a line that is parallel to the identified coordinate axis, and that is half way between the polygon's high and low coordinates on the coordinate axis perpendicular to the identified axis. At this stage, the process also specifies (at 3110) the x-, y-, s-, and t-coordinate values of the two endpoints of the segment. These two endpoints are two points on the boundary of the polygon that are intersected by the defined segment (i.e., are intersected by a line that is parallel to the identified coordinate axis and that is half way between the polygon's high and low coordinates on the coordinate axis perpendicular to the identified axis). For instance, in the example illustrated in FIG. 32, the segment 3220 is identified to be parallel to the s-axis and half way between the high and low t-axis coordinates of the polygon 3205. The endpoints of the segment 3220 are points 3225 and 3230.

[0094] The process then identifies the 8 delta values, dxLO, dyLO, dsLO, dtLO, dxHI, dyHI, dsHI, and dtHI, associated with the style. Specifically, it computes (at 3115) the low delta x-value dxLO by subtracting the polygon's x-low value xPLO from the x-low value xLO of the segment. The process then computes (at 3120) the low delta y-value dyLO by subtracting the polygon's y-low value yPLO from the y-low value yLO of the segment. It then computes (at 3125) the low delta s-value dsLO by subtracting the polygon's s-low value sPLO from the s-low value sLO of the segment. The process next computes (at 3130) the low delta t-value dtLO by subtracting the polygon's t-low value tPLO from the t-low value tLO of the segment.

[0095] The process then computes (at 3135) the high delta x-value dxHI by subtracting the x-high value xHI of the segment from the polygon's x-high value xPHI. It next computes (at 3140) the high delta y-value dyHI by subtracting the y-high value yHI of the segment from the polygon's y-high value yPHI. The process then computes (at 3145) the high delta s-value dsHI by subtracting the s-high value SHI of the segment from the polygon's s-high value sPHI. It next computes (at 3150) the high delta t-value dtHI, by subtracting the t-high value tHI of the segment from the polygon's t-high value tPHI.

[0096] The computed delta values dxLO, dyLO, dsLO, dtLO, dxHI, dyHI, dsHI, and dtHI define a style, which can be stored and indexed in a database table, such as table 2915 of FIG. 29. This style can be indexed by a segment data structure (such as structure 2905 of FIG. 29) for the segment specified at 3110. The direction of this segment would be one of the two directions parallel to the coordinate axis identified at 3105. The length of this segment would be the distance between the two endpoints of the segment.

[0097] III. Operations

[0098] The octangle data structure can be used to perform numerous operations extremely efficiently. Several such operations are described below.

[0099] A. Intersection.

[0100] The intersection of two design-layout items T1 and T2 that have octangle data structure representations can be identified as follows. Each intersected item TN is represented by four low values xLOTN, yLOTN, sLOTN, tLOTN and four high values sHITN, yHITN, sHITN, and tHITN. The intersection of these items is a design-layout item I (referred to below as an “intersection item”) that can be represented by four low values xLOI, yLOI, sLOI, tLOI, and four high values xHII, yHII, SHII and tHII.

[0101] To identify the intersection of two design-layout items T1 and T2, some embodiments first identify an initial set of low and high values (xLOI1, yLOI1, sLOI1, tLOI1, xHII1, yHII1, sHII1 and tHII1). In this initial set, the low values of the intersection item are initially defined as the maximum of the corresponding pair of low values of the intersected items, and the high values of the intersection item are initially defined as the minimum of the corresponding pair of high values of the intersected items. Specifically, the following equations define the initial low and high values.

x

LO

I1
=max(xLOT2 and xLOT2),

y

LO

I1
=max(yLOT1 and yLOT2),

sLOI1=max(sLOT1 and sLOT2),

t

LO

I1
=max(tLOT1 and tLOT2)

xHII1=min(xHIT1 and xHIT2),

y

HI

I1
=min(yHIT1 and yHIT2),

s

HI

I1
=min(sHIT1 and sHIT2),

t

HI

I1
=min(tHIT1 and tHIT2).

[0102] One or more of the identified, initial low and high values of the intersection item might specify a superfluous half-plane that is not necessary for defining the polygon at the intersection of the two design-layout items. Hence, after identifying the initial low and high values (xLOI1, yLOI1, sLOI1, tLOI1, xHII1, yHII1, sHII1, and tHII1) of the intersection item, some embodiments perform a canonicalization operation that might modify the identified low and high values of the intersection item. This canonicalization ensures that any superfluous half plane abuts one of the vertices of the polygon that defines the intersection of the two design-layout items. Eight low and high values (xLOI, yLOI, sLOI, tLOI, xHII, yHII, sHII, and tHII) remain after the canonicalization operation, and these eight values define the intersection item. One manner for performing a canonicalization operation is described below at the end of this section.

[0103] After canonicalizing the identified low and high values of the intersection item, some embodiment perform a check to ensure that the canonicalized, low and high values (xLOI, yLOI, sLOI, tLOI, xHII, yHII, sHII, and tHII) specify a valid polygon. This check compares the low and high values for each axis to ensure that high value is equal to or greater than the low value. Hence, for the intersection to be valid, the following conditions must be met:

xHII≧xLOI,

yHII≧yLOI,

sHII≧sLOI, and

t

HI

I
≧tLOI.

[0104]
FIG. 33 illustrates an example of identifying the intersection of a triangle 3305 and an octagon 3310. In this example, the triangle is specified by eight values xLOT, yLOT, sLOT, tLOT, xHIT, yHII, sHIT, and tHIT, and the octagon is specified by eight values xLOO, yLOO, sLOO, tLOO, xHIO, yHIO, sHIO, and tHIO. In this example, xLOO is larger than xLOT, tLOO is larger than yLOT, sLOO is larger than sLOT, tLOT is larger than tLOT, xHIO is smaller than xHIT, yHIT is smaller than yHIO, sHIT is smaller than SHIO, and tHIT is smaller than tHIO. Accordingly, as shown in FIG. 33, the initial set of half-planes that are identified for the intersection of the triangle 3305 and the octagon 3310 are xLOO, yLOO, sLOO, tLOO, xHIO, yHIT, sHIT, and tHIT. In this initial set, the x-hi value (xHIO) defines a superfluous half-plane 3330 that does not abut any vertex of the polygon 3320, which defines the intersection between the triangle 3305 and octagon 3310. After defining the initial set of values xLOO, yLOO, sLOO, tLOO, xHIO, yHIT, sHIT, and tHIT, a canonicalization operation is performed on this set of values. This canonicalization operation replaces the x-hi value (xHIO) in this set with a canonical x-hi value (xHIC). The canonical x-hi value (xHIc) defines a superfluous half-plane 3325 that abuts a vertex 3315 of the polygon 3320.

[0105] B. Union

[0106] The union of two design-layout items T1 and T2 that have octangle data structure representations can be identified as follows. The union of items T1 and T2 is a design-layout item U (referred to below as the union) that can be represented by four low values xLOU, yLOU, sLOU, tLOU and four high values xHIU, yHIU, sHIU, and tHIU. To identify the union of items T1 and T2, the low values of the union are identified as the minimum of the corresponding pair of low values of the merged items T1 and T2, and the high values of the union are identified as the maximum of the corresponding pair of high values of the merged items. Specifically, (1) the xLOU of the union is the minimum of the xLOT1 and xLOT2 of the items T1 and T2, (2) the yLOU of the union is the minimum of the yLOT1 and yLOT2 of the items T1 and T2, (3) the sLOU of the union is the minimum of the sLOT1 and sLOT2 of the items T1 and T2, (4) the tLOU of the union is the minimum of the tLOT1 and tLOT2 of the items T1 and T2, (5) the xHI of the union is the maximum of the xHIT1 and xHIT2 of the items T1 and T2, (6) the yHIU of the union is the maximum of the yHIT1 and yHIT2 of the items T1 and T2, (7) the SHIU of the union is the maximum of the SHIT1 and SHIT2 of the items T1 and T2, and (8) the tHIU of the union is the maximum of the tHIT1 and tHIT2 of the items T1 and T2. The union of two items does not yield an invalid data item. Hence, no validity check is necessary for a union operation.

[0107]
FIG. 34 illustrates an example of identifying a union of a rectangle 3405 and an octagon 3410. In this example, the rectangle is specified by eight values xLOR, yLOR, sLOR, tLOR, xHIR, yHIR, sHIR, and tHIR, and the octagon is specified by eight values xLOR, yLOR, sLOR, tLOR, xHIR, yHIO, sHIO, and tHIO. In this example, xLOR is smaller than xLOO, yLOR is smaller than yLOO, sLOR is smaller than sLOO, tLOR is smaller than tLOO, xHIO is larger than xHIR, yHIO is larger than yHIR, sHIO is larger than sHIR, and tHIR is larger than tHIO. Accordingly, as shown in FIG. 34, the union of the rectangle 3405 and the octagon 3410 is a polygon 3415 that can be represented by the eight values xLOR, yLOR, sLOR, tLOR, xHIO, yHIO, sHIO, and tHIR.

[0108] C. Expansion

[0109] It is often necessary to identify an expansion of an item in the design layout. For instance, an item that represents a pin or obstacle is often expanded in order to identify spacing constraints near the pin or obstacle. As described above, FIG. 23A illustrates an expansion of an item. In this example, the item is expanded by about 10 Manhattan units in the Manhattan directions and 15 diagonal units in the diagonal directions. As mentioned above, one diagonal unit is about

\frac{\sqrt{2}}{2}

[0110] times a Manhattan unit. Accordingly, 15 diagonal units in the diagonal directions is approximately equal to 10 Manhattan units in the Manhattan directions.

[0111] The expansion of an item T (referred to as the expanded item) in the design layout leads to another item E (referred to as the expansion item) in the layout. An item T can be expanded in a variety of ways. For instance, it can be expanded by specifying a unique expansion delta for each of its eight half-plane values, or by specifying one expansion delta along each of the four axes x, y, s, and t, or by specifying one expansion delta along the Manhattan directions and one expansion delta along the diagonal directions. In the discussion below, eight different expansion deltas dxLO, dyLO, dsLO, dtLO, dxHI, dyHI, dsHI, and dtHI are mentioned. One of ordinary skill, however, will realize that some or all of these deltas might specify the same value.

[0112] An expansion item E can be represented by four low values xLOE, yLOE, sLOE, tLOE and four high values xHIE, yHIE, sHIE, and tHIE. To identify the expansion of a design-layout item T, some embodiments first identify an initial set of low and high values (xLOE1, yLOE1, sLOE1, tLOE1, xHIE1, yHIE1, sHIE1, and tHIE1). In this initial set, each low value of the expansion item is initially specified as the corresponding low value of the expanded item minus the corresponding low value delta, while each high value of the expansion item is initially specified as the corresponding high value of the expanded item plus the corresponding high value delta. In other words, the eight half-plane values of the expansion item are initially computed as follows:

x

LO

E1

=x

LO

T

−dx

LO
;

y

LO

E1

=y

LO

T

dy

LO

T
;

sLOE1=sLOT−dsLOT;

t

LO

E1

=t

LO

T

−dt

LO

T
;

xHIE1=xHIT+dxHIT;

y

HI

E1

=y

HI

T

+dy

HI

T
;

sHIE1=sHIT+dsHIT; and

t

HI

E1

=t

HI

T

+dt

HI

T
.

[0113] After computing these eight values, some embodiments perform a canonicalization operation that might modify the identified initial expansion values (xLOE1, yLOE1, sLOE1, tLOE1, xHIE1, yHIE1, sHIE1, and tHIE1). This canonicalization operation ensures that any superfluous half plane abuts one of the vertices of the polygon that defines the expansion item. Eight low and high values (xLOE, yLOE, sLOE, tLOE, xHIE, yHIE, sHIE, and tHIE) remain after the canonicalization operation, and these eight values define the expansion item. One manner for performing the canonicalization is described at the end of this section.

[0114] After canonicalizing the identified low and high values of the expansion item, some embodiment perform a check to ensure that the canonicalized, low and high values specify a valid polygon. This check compares the low and high values for each axis to ensure that high value is equal to or greater than the low value. Hence, for the expansion to be valid, the following conditions must be met after the canonicalization:

x

HI

E

≧x

LO

E
,

yHIE≧yLOE,

sHIE≧sLOE, and

tHIE≧tLOE.

[0115] D. Reduction

[0116] The reduction of an item T (referred to as the reduced item) in the design layout leads to another item R (referred to as the reduction item) in the layout. An item T can be reduced in a variety of ways. For instance, it can be reduced by specifying a unique reduction delta for each of its eight half-plane values, or by specifying one reduction delta along each of the four axes x, y, s, and t, or by specifying one reduction delta along the Manhattan directions and one reduction delta along the diagonal directions. In the discussion below, eight different reduction deltas dxLO, dyLO, dsLO, dtLO, dxHI, dyHI, dsHI, and dtHI are mentioned. One of ordinary skill, however, will realize that some or all of these deltas might specify the same value.

[0117] A reduction item R can be represented by four low values xLOR, yLOR, sLOR, tLOR and four high values xHIR, yHIR, sHIR, and tHIR. To identify the reduction of a design-layout item T, some embodiments first identify an initial set of low and high values (xLOR1, yLOR1, sLOR1, tLOR1, xHIR1, yHIR1, sHIR1, and tHIR1). In this initial set, each low value of the reduction item is specified as the corresponding low value of the reduced item plus the corresponding low value delta, while each high value of the reduction item is specified as the corresponding high value of the reduced item minus the corresponding high value delta. In other words, the eight half-plane values of the reduction item are initially computed as follows:

xLOR1=xLOT+dxLOT;

y

LO

R1

=y

LO

T

+dY

LO

T
;

s

LO

R1

=s

LO

T

+ds

LO

T
;

t

LO

R1

=t

LO

T

+dt

LO

T
;

xHIR1=xHIT−dxHIT;

yHIR1=yHIT−dYHIT;

sHIR1=sHIT−dsHIT; and

tHIR1=tHIT−dtHIT.

[0118] After computing these eight values, some embodiments perform a canonicalization operation that might modify the initial values (xLOR1, yLOR1, sLOR1, tLOR1, xHIR1, yHIR1, sHIR1, and tHIR1) of the reduction item. This canonicalization operation ensures that any superfluous half plane abuts one of the vertices of the polygon that defines the reduction item. Eight low and high values (xLOR, yLOR, sLOR, tLOR, xHIR, yHIR, sHIR, and tHIR) remain after the canonicalization operation, and these eight values define the reduction item. One manner for performing canonicalization will be described at the end of this section.

[0119] After canonicalizing the identified low and high values of the reduction item, some embodiment perform a check to ensure that the canonicalized, low and high values specify a valid polygon. This check compares the low and high values for each axis to ensure that high value is equal to or greater than the low value. Hence, for the reduction to be valid, the following conditions must be met after the canonicalization:

xHIR≧xLOR,

yHIR≧yLOR,

sHIR≧sLOR, and

tHIR≧tLOR.

[0120]
FIG. 35 illustrates a portion of a polygon 3510 that is being reduced. In this example, the polygon 3510 is reduced by 10 Manhattan unit in the Manhattan directions and 15 diagonal units in the diagonal direction. The reduction operation initially identifies four low values and four high values for the reduction item. Each low value for the reduction item is initially specified as the corresponding low value of the reduced item plus the corresponding low value delta, and each high value for the reduction item is initially specified as the corresponding high value of the reduced item minus the corresponding high value delta. FIG. 35 illustrates three of the high values (xHIR, yHIR, sHIR) that are initially computed for the reduction of the polygon 3510.

[0121] In this initial set, the s-hi value (sHIR) defines a superfluous half-plane that does not abut any vertex of the polygon 3505, which defines the reduction of the polygon 3510. Hence, after defining the initial set of values, a canonicalization operation is performed on this set of values. This canonicalization operation replaces the s-hi value (sHIR) in this set with a canonical s-hi value (sHIC). The canonical s-hi value (sHIC) defines a superfluous half-plane 3515 that abuts a vertex 3520 of the polygon 3505, which defines the reduction of the polygon 3510.

[0122] E. Bounding Polygon

[0123] EDA tools often have to identify a bounding polygon or bounding box that encloses one or more items in the design layout. The octangle data structure can be used to specify such bounding polygons and boxes very quickly. As described above for the processes 2400 and 2500, all that needs to be done to identify a bounding polygon is to identify eight minimum and maximum x-, y-, s-, and t-values of the vertices of the set of items S for which the bounding polygon is identified. The eight minimum and maximum values specify eight half-planes that when intersected define an octilinear bounding polygon. From the identified bounding polygon, the Manhattan and rotated bounding boxes for the set of item S can then be quickly identified.

[0124]
FIG. 36 illustrates an example of constructing an octilinear bounding polygon for three design-layout items, which are a rectangle 3605, a triangle 3610, and a square 3615. In this example, the rectangle is specified by eight values xLOR, yLOR, sLOR, tLOR, xHIR, yHIR, sHIR, and tHIR, the triangle is specified by eight values xLOT, yLOT, sLOT, tLOT, xHIT, yHIT, sHIT, and tHIT, and the square is specified by eight values xLOS, yLOS, sLOS, tLOS, xHIS, yHIS, sHIS, and tHIS. In this example, xLOR is smaller than xLOS and xLOT, yLOS is smaller than yLOR and yLOT, sLOS is smaller than sLOT and sLOR, tLOT is smaller than tLOS and tLOR, xHIT is larger than xHIR and xHIS, yHIR is larger than yHIT and yLOS, sHIT is larger than sHIR and sHIS, and tHIR is larger than tHIS and tHIT. Accordingly, as shown in FIG. 36, the octilinear bounding polygon of the rectangle 3605, the triangle 3610, and the square 3615 is the polygon 3620 that can be represented by the eight values xLOR, yLOS, sLOS, tLOT, xHIT, yHIR, sHIT, and tHIR. As further shown in this figure, the Manhattan bounding box for these three items can be quickly identified as the rectangle 3625 defined by the low and high Manhattan values xLOR, yLOS, xHIT, and yHIR of the octilinear polygon 3620. Similarly, the rotated bounding box for these three items can be quickly identified as the rotated rectangle 3630 defined by the low and high s- and t-values sLOS, tLOT, sHIT, and tHIR of the octilinear polygon 3620.

[0125] F. Circumference

[0126] To compute the circumferences P of an octangle that is defined by a octangle data structure that specifies the eight values xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI, some embodiments use the following equation:

P
=(2−{square root}{square root over (2)})*(sHI+tHI−sLO−tLO)+2*({square root}{square root over (2)}−1)*(xHI+yHI−xLO−yLO).

[0127] This equation is derived by summing the following eight equations that express the length of each of the eight potential sides of the octangle.

Length of the vertical right side of the octangle=sHI−xHI−tLO−xHI.

Length of the horizontal top side of the octangle=sHI+tHI−2*yHI.

Length of the vertical left side of the octangle=tHI−sLO+2*xLO.

Length of the horizontal down side of the octangle=−tLO−sLO+2*yLO.

Length of diagonal side in the first diagonal quadrant=sHI*{square root}{square root over (2)}−(sHI−xHI)*{square root}{square root over (2)}−(sHI−yHI)*{square root}{square root over (2)}={square root}{square root over (2)}*(xHI+yHI−sHI)

Length of diagonal side in the second diagonal quadrant={square root}{square root over (2)}*(−xLO+yHI−tHI)

Length of diagonal side in the third diagonal quadrant={square root}{square root over (2)}*(−xLO−yLO+sLO)

Length of diagonal side in the fourth diagonal quadrant={square root}{square root over (2)}*(xHI−yLO+tLO)

[0128] G. Area

[0129] To compute the area A of an octangle that is defined by a octangle data structure that specifies the eight values xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI, some embodiments use the following equation:

\begin{matrix} A = (x_{HI} - x_{LO}) * (y_{HI} - y_{LO}) - (x_{HI} + y_{HI} - s_{HI}) * (x_{HI} + y_{HI} - s_{HI}) / 2 - \\ (- x_{LO} + y_{HI} - t_{HI}) * (- x_{LO} + y_{HI} - t_{HI}) / 2 - (- x_{LO} - y_{LO} + s_{LO}) * \\ (- x_{LO} - y_{LO} + s_{LO}) / 2 - (x_{HI} - y_{LO} + t_{LO}) * (x_{HI} - y_{LO} + t_{LO}) / 2 \end{matrix}

[0130] This equation is derived by first computing the area of a Manhattan bounding box of the octangle. At each of the Manhattan bounding box's corners, this box might extend beyond the octangle. Each such extension would be a triangle. Hence, after computing the area of the Manhattan bounding box, the area of the octangle can be expressed as the area of its Manhattan bounding box minus the areas of the box's potential triangular extensions beyond the octangle.

[0131] H. Canonicalization

[0132] As mentioned above, a set of values (xLO1, yLO1, sLO1, tLO1, xHI1, tHI1, sHI1, and tHI1) that define an initial octangle data structure might include one or more values that specify one or more superfluous half planes. In fact, an initial value might specify a superfluous half plane that does not abut a vertex of the shape that is defined by the octangle data structure. Hence, as mentioned above, some embodiments at times perform a canonicalization operation on the initial set of values (xLO1, yLO1, sLO1, tLO1, xHI1, yHI1, sHI1, and tHI1) to produce a canonical set of values (xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI), where each value in this canonical set either specifies a non-superfluous half plane or a superfluous half plane that abuts at least a vertex of the shape defined by the set. In some cases, the values in the canonical set are identical to their corresponding values in the initial set. In other cases, however, one or more values in the canonical set differ from their corresponding values in the initial set.

[0133] Table 1 below lists a set of equations that some embodiments use to define the canonical set of values (xLO, yLO, sLO, tLO, xHI, yHI, sHI, and tHI) from the initial set of values (xLO1, yLO1, sLO1, tLO1, xHI1, yHI1, sHI1, and tHI1). The set of equations specified in this table use parameters x′, y′ and s′, t′, which can be different for different canonical values. Accordingly, Table 1 provides the definition of the parameters x′, y′ and s′, t′ for different canonical values. These parameters are defined at the intersection of two edges of two half planes. The intersection of two edges of two half planes are defined by identifying the intersection of two lines that are specified by the two values that define the two half planes.

[0134]
FIG. 37 provides an example of the intersection of two lines that represent two edges of two half plane. Specifically, this figure illustrates an intersection point 3705 between line 3710 that is an edge of a half plane 3715 specified by the value sLO and line 3720 that is an edge of a half plane 3725 specified by the value xHI. To identify such an intersection point, some embodiments first define an expression for each of the lines in the same coordinate system (e.g., in the example illustrated in FIG. 37, some embodiments define an equation in the x-y coordinate system for the line 3710 and use this equation in conjunction with the equation x=xHI that defines the line 3720). After expressing both lines in the same coordinate system, these embodiments then identify the intersection of the two lines by applying known line-intersection techniques.

1TABLE 1CanonicalValuex′, y′ and s′, t′Equations to Use to Compute Canonical ValuexLOx′ and y′ are the x andIf y′ <yLO 1, then xLO = max (xLO1, x coordinate aty coordinates at theintersection of the lines specified by yLO1 and tHI1)intersection of theIf y′ <yHI 1, then xLO = max (xLO1, x coordinate atlines specified by sLO1intersection of the lines specified by sLO1 and yHI1)and tHI1If yLO1 ≦ y′ ≦ yHI1, then xLO = max (xLO1, x′).yLOx′ and y′ are the x andIf x′ < xLO1, then yLO = max (yLO1, y coordinate aty coordinates at theintersection of the lines specified by xLO1 and tLO1)intersection of theIf x′ > xHI1, then yLO = max (yLO1, y coordinate atlines specified by sLO1intersection of the lines specified by sLO1 and xHI1).and tLO1If xLO1 ≦ x′ ≦ xHI1, then yLO = max (yLO1, y′).sLOs′ and t′ are the s and tIf t′ < tLO1, then sLO = max (sLO1, s coordinate atcoordinates at theintersection of the lines specified by yLO1 and tLO1)intersection of theIf t′ > tHI1, then sLO = max (sLO1, s coordinate atlines specified by xLO1intersection of the lines specified by xLO1 and tHI1).and yLO1If tLO1 ≦ t′ ≦ tHI1, then sLO = max (sLO1, s′).tLOs′ and t′ are the s and tIf s′ < sLO1, then tLO = max (tLO1, t coordinate atcoordinates at theintersection of the lines specified by xHI1 and sLO1)intersection of theIf s′ > sHI1, then tLO = max (tLO1, t coordinate atlines specified by yLO1intersection of the lines specified by yLO1 and sHI1).and xHI1If sLO1 ≦ s′ ≦ sHI1, then tLO = max (tLO1, t′).xHIx′ and y′ are the x andIf y′ < yLO1, then xHI = min (xHI1, x coordinate aty coordinates at theintersection of the lines specified by yLO1 and sHI1)intersection of theIf y′ > yHI1, then xHI = min (xHI1, x coordinate atlines specified by sHI1intersection of the lines specified by tLO1 and yHI1).and tLO1If yLO1 ≦ y′ ≦ yHI1, then xHI = min (xHI1, x′).yHIx′ and y′ are the x andIf x′ < xLO1, then yHI = min (yHI1, y coordinate aty coordinates at theintersection of the lines specified by xLO1 and sHI1)intersection of theIf x′ > xHI1, then yHI = min (yHI1, y coordinate atlines specified by sHI1intersection of the lines specified by tHI1 and xHI1).and tHI1If xLO1 ≦ x′ ≦ xHI1, then yHI = min (yHI1, y′).sHIs′ and t′ are the s and tIf t′ < tLO′, then sHI = min (sHI1, s coordinate atcoordinates at theintersection of the lines specified by tLO1 and yHI1)intersection of theIf t′ > tHI1, then sHI = min (sHI1, s coordinate atlines specified by xHI1intersection of the lines specified by xHI1 and tHI1).and yHI1If tLO1 ≦ t′ ≦ tHI1, then sHI = min (sHI1, s′).tHIs′ and t′ are the s and tIf s′ < sLO1, then tHI = min (tHI1, t coordinate atcoordinates at theintersection of the lines specified by sLO1 and yHI1)intersection of theIf s′ > sHI1, then tHI = min (tHI1, t coordinate atlines specified by xLO1intersection of the lines specified by xLO1 and sHI1).and yHI1If sLO1 ≦ s′ ≦ sHI1, then tHI = min (tHI1, t′).

[0135] IV. Computer System

[0136]
FIG. 38 conceptually illustrates a computer system with which some embodiments of the invention are implemented. Computer system 3800 includes a bus 3805, a processor 3810, a system memory 3815, a read-only memory 3820, a permanent storage device 3825, input devices 3830, and output devices 3835.

[0137] The bus 3805 collectively represents all system, peripheral, and chipset buses that communicatively connect the numerous internal devices of the computer system 3800. For instance, the bus 3805 communicatively connects the processor 3810 with the read-only memory 3820, the system memory 3815, and the permanent storage device 3825.

[0138] From these various memory units, the processor 3810 retrieves instructions to execute and data to process in order to execute the processes of the invention. The read-only-memory (ROM) 3820 stores static data and instructions that are needed by the processor 3810 and other modules of the computer system. The permanent storage device 3825, on the other hand, is read-and-write memory device. This device is a non-volatile memory unit that stores instruction and data even when the computer system 3800 is off. Some embodiments of the invention use a mass-storage device (such as a magnetic or optical disk and its corresponding disk drive) as the permanent storage device 3825. Other embodiments use a removable storage device (such as a floppy disk or zip® disk, and its corresponding disk drive) as the permanent storage device. Like the permanent storage device 3825, the system memory 3815 is a read-and-write memory device. However, unlike storage device 3825, the system memory is a volatile read-and-write memory, such as a random access memory. The system memory stores some of the instructions and data that the processor needs at runtime. In some embodiments, the invention's processes are stored in the system memory 3815, the permanent storage device 3825, and/or the read-only memory 3820.

[0139] The bus 3805 also connects to the input and output devices 3830 and 3835. The input devices enable the user to communicate information and select commands to the computer system. The input devices 3830 include alphanumeric keyboards and cursor-controllers. The output devices 3835 display images generated by the computer system. For instance, these devices display IC design layouts. The output devices include printers and display devices, such as cathode ray tubes (CRT) or liquid crystal displays (LCD).

[0140] Finally, as shown in FIG. 38, bus 3805 also couples computer 3800 to a network 3865 through a network adapter (not shown). 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 or all of the components of computer system 3800 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.

[0141] While the invention has been described with reference to numerous specific details, one of ordinary skill in the art will recognize that the invention can be embodied in other specific forms without departing from the spirit of the invention. For instance, some embodiments might also include a layer attribute in the octangle data structure 205 of FIG. 2, while other embodiments might specify the layer information in some other manner.

[0142] Also, several of the above-described embodiments use the octangle data structure 205, which is useful to represent design-layout items with horizontal, vertical, and ±45° directions. Other embodiments, however, might use other types of data structures that are optimized for other types of design-layout items.

[0143] For example, some embodiments define routes according to a hexagonal wiring model that specifies three routing directions that are offset from each other by 60°. For such a wiring model, some embodiments use a hexangle data structure 3900 that is illustrated in FIG. 39. This data structure stores eight values that are defined with respect to a hexagonal coordinate system 4000 illustrated in FIG. 40. As shown in FIG. 40, the hexagonal coordinate system includes a Manhattan coordinate system and a rotated coordinate system. The rotated coordinate system includes u- and v-axes. The u-axis is at a 60° clockwise rotation from the y-axis, while the v-axis is at a 60° counterclockwise rotation from the y-axis. The hexangle data structure represents each convex geometric shape in terms of eight values, xLO, yLO, uLO, vLO, xHI, yHI, uHI, and vHI. These eight values define eight half planes that bound a design-layout item in the Manhattan and rotated coordinate systems. The eight values are useful in representing design-layout items (such as pins) that have sides in the Manhattan and ±30° directions (where 30° is defined with respect to the Manhattan coordinate system). In addition, the y-, u-, and v-values are useful in representing routes that have segments parallel to the y-, u-, and v-axis.

[0144] In some cases, the hexagonal wiring model specifies the x-axis as one of the routing directions. In these cases, the three routing directions are 0°, 60°, and −60°. In these cases, the u- and v-values of the hexangle data structure 3900 would be defined by reference to the coordinate system of FIG. 41. This coordinate system is similar to the coordinate system 4000 of FIG. 40, except that the u′- and v′-axes in FIG. 41 are rotated by 600 with respect to the x-axis.

[0145] Alternatively, instead of specifying a data structure that is optimized for one of the two potential hexagonal wiring models mentioned above, some embodiments use a symmetric-hexangle data structure 4200 illustrated in FIG. 42. This data structure is defined for the coordinate system of FIG. 43, which includes a Manhattan coordinate system, a 30°-rotated coordinate system, and a 60°-rotated coordinate system. The 30°-rotated coordinate system includes two axes, u- and v-axes, that are rotated by 30° with respect to the x-axis (i.e., are rotated by 60° with respect to the y-axis). The 60°-rotated coordinate system includes two axes, u′- and v′-axes, that are rotated by 60° with respect to the x-axis.

[0146] The symmetrical hexangle data structure 4200 represents each convex geometric shape in terms of twelve values, xLO, yLO, uLO, vLO, u′LO, v′LO, xHI, yHI, uHI, vHI, u′HI, and v′HI. These twelve values define twelve half planes that bound a design-layout item in the three sub-coordinate systems of the coordinate system of FIG. 43. The twelve values are useful in representing design-layout items (such as pins or routes) that might have sides in the Manhattan, ±30°, and ±60° directions. Thus, one of ordinary skill in the art would understand that the invention is not to be limited by the foregoing illustrative details, but rather is to be defined by the appended claims.

Method and apparatus for representing items in a design layout

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

US Classifications

International Classifications

Abstract

Description

Claims

CLAIM OF BENEFIT

Provisional Applications (1)