JPH10283390A - Data compressing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make the efficiency of data handling high and to improve the throughput by representing data by using an index for discriminating the shape of a figure and start-point coordinates of the figure in a specific area. SOLUTION: When LSI layout data is inputted (S101) and the input data is a cell name (S106), cell name information is retrieved, and additionally registered (S107), and a cell name index is registered as instance information (S108). When the data is a figure or cell quotation (S109), shape information is retrieved or additionally registered (S110) and the figure or an offset quantity and a shape index are registered as instance information (S111). When the figure has array information (S112), the array information is retrieved or additionally registered (S113) and an array index is registered is instance information (S114). After the registering processes, data are outputted after being compressed by using the cell name, shape, array index, and start point coordinates of the figure.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data compression method for reducing a data size, and particularly to an LSI (Laser
The present invention relates to a data compression method for layout data.

[0002]

2. Description of the Related Art In general, in an LSI layout design process, the arrangement of elements and wirings in a chip is often represented by LSI layout data in a format called a stream format (or GDSII format). By the way, the arrangement of elements and wiring in a chip is
When this is represented by a figure shape, there are many figures represented by rectangles, and most of the figures are parallel to the X and Y axes.

Therefore, in the conventional LSI layout data, these graphic shapes are represented by the coordinate values of the vertices of each graphic shape. For example, when expressing two rectangles G and H shown in FIG.
In the layout data, G); (0, 0) (0, 2) (1, 2) (1, 0) (0, 0) H); (2, 3) (2, 5) (3, 5) (3, 3) (2, 3) are represented by a coordinate sequence.

[0004]

However, in the LSI layout data using the above-described coordinate sequence, there is a problem that the efficiency of data processing is reduced when the density of the LSI is increased. That is, the conventional LS
In the I layout data, as the density of LSI increases,
The data size becomes enormous, and much time is required for data handling such as input / output of data and data retrieval from a data structure. Further, when data processing is performed by a computer (for example, a CAD device or the like) that employs a virtual memory system that is currently common, the amount of memory used becomes about twice the mounted memory size due to the enormous data size. If this occurs, a phenomenon called thrashing occurs, and the processing efficiency is extremely reduced.

Accordingly, the present invention provides a data compression method capable of improving the data handling efficiency by efficiently reducing the data size, and improving the processing efficiency including avoiding thrashing due to the memory saving of the computer. The purpose is to provide.

[0006]

SUMMARY OF THE INVENTION The present invention is a data compression method devised to achieve the above object. That is, the data compression method of the present invention is a data compression method for data indicating a graphic arrangement in a predetermined area,
An index for identifying the shape is given for each shape of the figure, and the data is expressed using the index and the coordinates of the starting point position of the figure in the predetermined area.

According to the data compression method according to the above procedure,
In the data indicating the graphic arrangement in the predetermined area, if there are figures of the same shape arranged at different positions,
Give these figures the same index,
These figures are represented using the index and the starting position coordinates of each figure. Therefore, for example, the data size is reduced as compared with a case where each figure is represented by a coordinate sequence of vertex coordinates.

In the data compression method of the present invention, when the index is given, a shape of a figure to which the index is given may be defined by using a displacement between vertex coordinates in the shape.

Thus, the shape of each figure is represented by the amount of displacement between the coordinates of the vertices. Only the value of either X or Y is required for each vertex. In addition, since the amount is the amount of displacement, the value itself becomes small and the amount of information is reduced, so that the data size is reduced.

Further, in the data compression method according to the present invention, when the predetermined area is composed of a plurality of layers, different indices are given to different figures even if the figure has the same shape and the arranged layers are different. Good.

[0011] In this way, the layers can be virtually distinguished from each other in the form of graphic information, so that the instructions regarding the arrangement can be handled in the same manner as the cell arrangement described later.

Further, the data compression method of the present invention regards the arrangement of cells as a set of a plurality of figures as one figure shape, and for each cell, and its rotation information (for example, the rotation angle, the presence or absence of mirroring, , Magnification, etc.).

[0013] With this, even for a cell citation frequently used in LSI layout data, for example, a cell including rotation is regarded as one graphic shape and an index is given, so that each cell citation is referred to as a cell name and each rotation information. The data size is reduced as compared with the case of expressing by.

According to the data compression method of the present invention, when a plurality of figures having the same shape are arranged in a predetermined pattern in the predetermined area, an index is given to the predetermined pattern, and the index is assigned to the same index. The data may be represented by combining an index given to the shape.

Thus, when there are a plurality of figures arranged in a certain pattern, these figures are assigned to the index corresponding to the shape, the index given to the certain pattern, and the index of the certain pattern. This can be represented by the start position coordinates, and the data size can be further reduced.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A data compression method according to the present invention will be described below. Here, a data compression method for LSI layout data will be described as an example.

First, an outline of the data compression method according to the present embodiment will be described. According to this data compression method, an index (hereinafter, referred to as INDEX) for identifying the shape of each figure showing elements and wirings in a chip is given,
By expressing the LSI layout data indicating the arrangement of the elements and wirings using the INDEX and the coordinates of the starting point of each figure, the data size of the LSI layout data is reduced.

For example, when the two rectangles G and H shown in FIG. 8 are expressed, the shape of each rectangle is the same in the LSI layout data compressed by this data compression method. 1) Shape type; rectangle, W (width) = 1, H (height) = 2, G); INDEX = # 1, (0, 0) H); INDEX = # 1, (2, 3) Express as follows. This is a result of focusing on the empirical fact that the appearance probability of a figure having the same shape is high in the LSI layout data. Here, a rectangle has been described as an example, but not only a rectangle but also an arbitrary polygon, a so-called path figure, or a L
All information in the SI layout data corresponds.

Next, details of the data compression method of the present embodiment will be described with reference to the drawings. FIG. 1 is a flowchart showing a procedure of a data compression method according to the present embodiment, and FIG. 2 is an explanatory diagram showing an outline of a data configuration of LSI layout data compressed by the data compression method according to the present embodiment. . However, the data compression method of the present embodiment uses an input / output interface (I / F) for inputting / outputting LSI layout data, a CPU (Cent
ral processing unit) and R
It is assumed that the program is executed by a computer having a work area such as an AM (Random Access Memory).

Here, the data structure of the LSI layout data compressed by the data compression method of the present embodiment will be described first with reference to FIG. As shown in the syntax section, the LSI layout data after compression includes header information located at the beginning and information on a set of graphics generally called cells after this header information, and is repeated as necessary. And cell information.

The header information includes information relating to the date of creation and the version number. Cell information starts with "bigin" in the syntax and includes "en
d ”, and consists of a cell name INDEX and instance information.

Among the cell information, the cell name INDEX is an INDEX registered in the cell name information including the cell name INDEX and the cell name. On the other hand, the instance information includes information for specifying the shape and arrangement position of the figure in the LSI layout data, such as an offset amount, shape INDEX, and array INDEX, which will be described in detail later. If the property information is added to the LSI layout data before compression, the property information may be held in the instance information.

In order to obtain LSI layout data having such a data structure, in the present embodiment, data compression is performed according to a procedure as shown in the flowchart of FIG. First, the input / output I / F of the computer
, The LSI layout data before compression is input (step 101; hereinafter, step is abbreviated as S). If the input data is header information (S102), the compression processing execution means registers the header information in the work area (S103).

If the input data is information indicating the start or end of a cell (S104), the information is registered in the work area (S105). If the input data is a cell name (S106), cell name information is searched or newly added and registered (S107), and a cell name INDEX indicating the cell name is registered as one of the instance information in the work area. (S108).

If the input data indicates a figure or cell citation (S109), the shape information is searched or newly added and registered (S110), and the offset amount indicating the cell citation including the figure or rotation information is obtained. And the shape INDEX are registered in the work area as one of the instance information (S111). Further, if the figure has array information (S112), the array information is searched or newly added and registered (S113), and an array INDEX indicating the array information is registered as one of the instance information in the work area. (S114).

These processes are repeated until the LSI layout data input to the input / output I / F becomes EOF (end of file) (S115). When it comes to EOF,
The compression processing execution means reconstructs various information registered in the work area in accordance with the above-described data configuration, and outputs it as compressed LSI layout data from the input / output I / F (S116).

Here, the data compression method described above is referred to as LS
A case where the present invention is applied to a specific figure shape in the I layout data will be described with reference to FIG. In the following description, it is assumed that the LSI layout data is composed of a plurality of layers (Layers; hereinafter, referred to as layers).

[Case of Rectangular] As shown in FIG.
For a rectangle existing in layer = 1, the compression processing execution means extracts W and H from the lower left vertex coordinates and the upper right vertex coordinates of the rectangle data, and obtains the shape information (FIG. 3B) in the work area. Search) in the same layer and in the same layer. If you already have the same rectangle,
Get the rectangular shape INDEX. On the other hand, if the same rectangle does not exist, the necessary information (layer, W,
H) is registered, a new shape INDEX is assigned, and then the shape INDEX is acquired. The acquired shape INDEX and the lower left coordinate value are registered as instance information in the work area. FIG. 3C briefly shows this. As described above, the lower left coordinate value, that is, the starting point position slip of the rectangle is registered as the offset amount. As a result, the rectangle is represented by the offset amount and the shape INDEX.

[In the case of a trapezoid parallel to the X axis (Xtrap trapezoid)] In the case of an Xtrap trapezoid as shown in FIG.
Of the two sides parallel to the axis, H, dx1 and dx2 are extracted from the side W having the smaller Y coordinate value, and there is an Xtrap trapezoid of the same layer and the same shape in the shape information in the work area. Search for no. If there is already the same Xtrap trapezoid, the shape INDEX of the Xtrap trapezoid is acquired. On the other hand, the same Xtr
ap If there is no trapezoid, register necessary information in the work area, assign a new shape INDEX, and then
To get. The acquired shape INDEX and the minimum coordinate value of W are registered as instance information in the work area. The minimum coordinate value of W is registered as an offset amount,
Let it be the starting point of the Xtrap trapezoid. As a result, the Xtrap trapezoid is represented by the offset amount and the shape INDEX.

[In the case of a trapezoid parallel to the Y axis (Ytrap trapezoid)] In the case of a Ytrap trapezoid as shown in FIG.
Of the two sides parallel to the axis, W, dy1 and dy2 are extracted from the side H having the smaller X coordinate value, and there is a Ytrap trapezoid of the same layer and the same shape in the shape information in the work area. Search for no. If the same Ytrap trapezoid already exists, the shape INDEX of the Ytrap trapezoid is acquired. On the other hand, the same Ytr
ap If there is no trapezoid, register necessary information in the work area, assign a new shape INDEX, and then
To get. The acquired shape INDEX and the minimum coordinate value of H are registered as instance information in the work area. The minimum coordinate value of H is registered as an offset amount,
The starting point of the Ytrap trapezoid. As a result, the Ytrap trapezoid is represented by the offset amount and the shape INDEX.

[In the case of a polygon whose sides are all parallel to the X-axis or Y-axis (polygon of xytype = 0)] In the case of a polygon of xytype = 0 as shown in FIG. As a result of the Y minimum search of the value and then the X minimum search, the corresponding vertex is set to “1”.
And sort the vertex coordinates clockwise. When such normalization is performed, the shape of this polygon is defined by the number of Xs and Ys (the number of vertices minus 1) starting from the vertex “1” and starting at the Y coordinate y2 of the vertex “2” and ending at yn. It can be represented by alternating coordinate strings.
Using such properties, extract layers and coordinate sequences,
The same layer and the shape information in the work area
Search for polygons of the same shape. If the same polygon already exists, get the shape INDEX of the polygon. On the other hand, if the same polygon does not exist, necessary information is registered in the work area, a new shape INDEX is assigned, and then the shape INDEX is acquired. Obtained shape INDEX and vertex “1”
Are registered as instance information in the work area. The coordinate value of the vertex “1” is registered as an offset amount, and is set as the starting point of the polygon. As a result, a polygon whose sides are all parallel to the XY axes is represented by the offset amount and the shape INDEX.

[In the case of a polygon having at least one side not parallel to the X axis or Y axis (polygon of xytype = 1)] In the case of a polygon of xytype = 1 as shown in FIG. As a result of the Y minimum search and then the X minimum search of the vertex coordinate values, the corresponding vertex is set to “1” and the vertex coordinates are sorted clockwise. After performing such normalization, the XY coordinates x2, y2 of the vertex “2”
(The number of vertices minus 1) XY coordinate strings are extracted, and it is searched whether or not the shape information in the work area includes a polygon having the same layer and the same shape. If the same polygon already exists, get the shape INDEX of the polygon. On the other hand, if the same polygon does not exist, the necessary information is registered in the work area and the shape INDEX is obtained. Acquired shape INDEX
Is registered as instance information in the work area. As a result, the polygon of xytype = 1 is represented by the offset amount and the shape INDEX.

[Case where all sides are parallel to the X-axis or Y-axis and the first side is a horizontal path (path of xytype = 0)] FIG.
In the case of a path of xytype = 0 as shown in (h), the minimum X vertex of the first side is set to “1”, and its coordinate value is set as the starting point (offset amount) of the figure in the instance information in the work area. register. Thereafter, the coordinate value of the vertex “1” is offset to (0, 0), and the other vertex coordinates are also offset. As a result, the shape of this path can be represented by an alternate coordinate sequence of X and Y starting from the X coordinate x2 of the vertex “2”. Then, it is searched whether or not the shape information in the work area includes a path having the same layer and the same shape. If there is already the same path, the shape of that path INDEX
To get. On the other hand, if the same path does not exist, the necessary information is registered in the work area and the shape INDEX is obtained. The acquired shape INDEX and the coordinate value of the vertex “1” are registered as instance information in the work area. As a result,
The path of xytype = 0 is represented by the offset amount and the shape INDEX.

[Case where all sides are parallel to the X-axis or Y-axis and the first side is a vertical path (path of xytype = 1)] FIG.
In the case of the path of xytype = 1 as shown in (i), the minimum vertex of Y on the first side is set to “1”, and its coordinate value is set as the starting point (offset amount) of the figure in the instance information in the work area. register. Thereafter, the coordinate value of the vertex “1” is offset to (0, 0), and the other vertex coordinates are also offset. As a result, the shape of this path can be represented by an alternate coordinate sequence of X and Y starting from the Y coordinate y2 of the vertex “2”. Then, it is searched whether or not the shape information in the work area includes a path having the same layer and the same shape. If there is already the same path, the shape of that path INDEX
To get. On the other hand, if the same path does not exist, the necessary information is registered in the work area and the shape INDEX is obtained. The acquired shape INDEX and the coordinate value of the vertex “1” are registered as instance information in the work area. As a result,
The path of xytype = 1 is represented by the offset amount and the shape INDEX.

[In the case of a path having at least one side that is not parallel to the X axis or Y axis (path of xytype = 2)] In the case of a path of xytype = 2 as shown in FIG. Is set to "1", and its coordinate value is registered as the starting point (offset amount) of the figure in the instance information in the work area. Thereafter, the coordinate value of the vertex “1” is offset to (0, 0), and the other vertex coordinates are also offset.
Then, an XY coordinate sequence starting from the vertex “2” is extracted, and the same layer and the same Width are added to the shape information in the work area.
And whether there is a path having the same shape. If the same path already exists, get the shape INDEX of that path. On the other hand, if the same path does not exist, the necessary information is registered in the work area and the shape INDEX is obtained. Acquired shape IN
DEX and the coordinate value of vertex “1” are registered as instance information in the work area. As a result, xytype =
The second path is represented by the offset amount and the shape INDEX.

[Cell Citation] In the case of a cell citation as shown in FIG. 3 (k), a cell citation coordinate value is first extracted, and the coordinate value is registered as an offset amount in instance information in the work area. I do. Next, the quoted cell name, rotation (Reflect, Angle, Mag), array (Nx, N
y, Dx, Dy), and search for cell name information, rotation information, and sequence information for each cell.
Get DEX, rotation INDEX, array INDEX.
Then, a search is performed to determine whether the same combination of the cell name INDEX and the rotation INDEX exists in the work area. If the same information already exists, the shape INDEX of the information is acquired. If the same information does not exist, necessary information is registered in the work area, the shape INDEX is newly assigned, and then the shape INDEX is acquired. Acquired shape INDEX and array INDEX
Is registered as instance information in the work area. As a result, the cell quotation is
And the array INDEX. Searching for cell name information, rotation information, and sequence information and obtaining INDEX are performed in the same manner as shape information.

As described above, according to the data compression method of the present embodiment, the shape of a figure (for example, a rectangle, trapezoid, polygon, or path) in LSI layout data is changed between vertex coordinates in each shape. They are defined using quantities, and the results are registered as instance information. Further, in this data compression method, when the LSI layout data is composed of a plurality of layers, even if the figure has the same shape, if the layers to be arranged are different, each has a different shape INDEX. I have. In other words, the layers are virtually distinguished from each other in the form of graphic information, so that the instruction regarding the arrangement can be handled in the same manner as the cell arrangement. Further, in this data compression method, even in the case of cell citation, the arrangement of cells is regarded as one figure shape, and a shape INDEX is assigned to each cell.

Here, a specific example of the LSI layout data after data compression by the above-described data compression method will be described. FIG. 4 is an example in which the LSI layout data after compression is dumped. Note that here, for simplicity of explanation, the figures are only rectangles and cell quotes,
Simple rotation and arrangement were used for cell citation. FIG. 5 is an example in which the LSI layout data shown in FIG. 4 is illustrated.

In the LSI layout data shown in FIG. 4, (A) in the instance information has a start point position of offset (3, 10) and 1 × 1 from the array INDEX = # 1.
, And a rectangle of W = 10 and H = 5 from the shape INDEX = # 1. When this is illustrated, a rectangle A shown in FIG. 5 is obtained. In (B) in the instance information of FIG. 4, the start point position is offset (15, 10), and the array INDEX =
From # 2, an array of 2 x 2 with an interval of 6 in both XY, shape INDEX =
From # 2, a rectangle of W = 5 and H = 5 is represented. When this is illustrated, a rectangle B shown in FIG. 5 is obtained. (C) in the instance information of FIG.
0), and the array INDEX = 1 × 1 array and shape from # 1
From INDEX = # 3, cell name INDEX # 2 (B), rotation INDEX # 1 (Reflect = 0, Angle = 0, Mag = 1.
0). This is illustrated in FIG.
Becomes the cell citation C shown in FIG. (D), (E), and (F) in the instance information in FIG. 4 are the same as those in (A).

Next, in addition to the above-described procedure of the data compression method, a case where repetitive information is extracted will be described with reference to the flowchart of FIG. The repetition information is information for defining an arrangement pattern when a plurality of figures having the same shape are arranged in a fixed pattern, similarly to the arrangement information in the case of cell quotation.

For example, when there are two or more instance information items having the same information except for the offset amount, the LSI layout data input to the input / output I / F is
When it becomes F (S215), subsequently, the compression processing execution means extracts repetitive information (S216). The repetition information includes the number of repetitions Nx in the X direction and the number of repetitions in the Y direction
Ny, X-direction repetition interval Dx, and Y-direction repetition interval Dy
For example, when only one figure is arranged, it is extracted as (1,1,0,0).

When the repetition information is extracted, the compression processing execution means determines the arrangement information in the work area (see FIG. 7A).
The same repetition information (Nx, N
(y, Dx, Dy). If the same repetition information already exists, get the array INDEX. On the other hand, if the same repetition information does not exist, the extracted repetition information is registered, and a new array INDEX is assigned.
DEX is acquired (S217). The obtained array INDEX is
It is registered as instance information in a predetermined format (S218). Then, the compression processing execution means reconfigures various information registered in the work area, and outputs the information as compressed LSI layout data from the input / output I / F (S219).

As described above, when there is two or more instance information in which all information other than the offset amount is the same, the instance information can be efficiently expressed by combining the shape INDEX and the array INDEX. it can. For example, when there is a figure which can be specified by the offset amount and the shape INDEX as shown in FIG.
When X and the array INDEX are combined, they can be expressed as shown in FIG.

As described above, the data compression method of the present embodiment focuses on the empirical fact that the appearance probability of a figure having the same shape is high in the LSI layout data.
If there are figures of the same shape placed at different positions in the layout data, the same
Give INDEX and convert these figures to their INDEX
And the starting point coordinates of each figure, that is, the offset amount. Therefore, by performing data compression on LSI layout data using this data compression method, the data size can be reduced as compared with the conventional case where each figure is represented by a coordinate sequence of vertex coordinates. . This makes it possible to suppress the data size from becoming enormous even if the density of the LSI is increased, and to spend a lot of time on data handling and reduce computer processing efficiency due to thrashing. Can be prevented.

In view of the characteristic that there are many figures represented by rectangles in the LSI layout data and most of the figures are parallel to the X and Y axes, the data compression method according to the present embodiment Is defined using the amount of displacement between the vertex coordinates in the shape (for example, W and H for a rectangle). Thus, compared to the conventional case where the shape of each figure is represented by a coordinate sequence of vertex coordinates, in the case of a figure composed of only line segments parallel to the XY axes, only the value of either X or Y is provided for each vertex. Get well. In addition, because of the amount of displacement, the value itself is reduced and the amount of information is reduced, so that the data size can be reduced.

In consideration of the characteristic that the LSI layout data is usually composed of a plurality of layers,
In the data compression method according to the present embodiment, different shapes INDEX are given to different figures even if they have the same shape if the layers to be arranged are different. This allows
By making the layer distinction virtually graphic shape information, the instruction regarding the arrangement can be handled in the same manner as the cell arrangement.

In view of the characteristic that cell layout is frequently used in LSI layout data, the data compression method according to the present embodiment regards the arrangement of cells, which are a set of a plurality of figures, as one figure shape. INDEX such as cell name INDEX and rotation INDEX is given every time. As a result, the arrangement of each cell can be handled in the same manner as one graphic shape, and as a result, the data size can be reduced as compared with the conventional case.

In view of the characteristic that the appearance probability of a figure having the same shape is high in the LSI layout data, the data compression method according to the present embodiment uses the case where a plurality of figures having the same shape are arranged in a fixed pattern. Provides the array INDEX for this fixed pattern. As a result, when a plurality of figures are arranged in a fixed pattern, these figures can be represented by the offset amount, the shape INDEX, and the array INDEX, and the data size can be further reduced. . This is the same in the case of cell quotation.

As a result of applying such a data compression method according to the present embodiment to an actual ASIC (application specific IC) circuit pattern of the 0.25 μm rule, the data size is 100 MB in the stream format (GDSII format). About 10,000
The data size of LSI layout data composed of various types of graphics and having a ratio of rectangular graphics to other graphics of 8: 2 could be reduced to about 40 MB. That is, in this case, a compression ratio of about 40% can be obtained by the data compression method of the present embodiment,
As a result, the time required for data handling is reduced from about 20 minutes before compression to about 8 minutes, and the memory usage of the computer can be reduced by about 60%.

Also, as a result of applying the data compression method according to the present embodiment to an actual SRAM circuit pattern of the 0.25 μm rule, the data size is 30 MB in the stream format (GDSII format) and approximately 20 MB.
The data size of LSI layout data composed of 000 types of graphics and having a ratio of rectangular graphics to other graphics of 6: 4 was reduced to about 10 MB. In other words, in this case, a compression rate of about 33% can be obtained by the data compression method of the present embodiment, whereby the time required for data handling is reduced from about 7 minutes before compression to about 3 minutes, and furthermore, the computer Memory savings of about 67%.

In the present embodiment, the data compression method for LSI layout data has been described as an example, but the present invention is not limited to this. For example,
Any data other than the LSI layout data can be applied as long as the data indicates a graphic arrangement in a predetermined area, such as image data indicating a graphic layout.

[0052]

As described above, in the data compression method of the present invention, the data indicating the graphic arrangement is represented by the index for each graphic shape and the starting position coordinates of each graphic. Therefore, by using this data compression method, the data size can be reduced as compared with the conventional case where each figure is represented by a coordinate sequence of vertex coordinates. As a result, it is possible to suppress the data size of the data indicating the graphic arrangement from becoming enormous, thereby improving the efficiency of data handling, and improving the processing efficiency including avoiding thrashing by conserving computer memory. This has the effect of becoming

[Brief description of the drawings]

FIG. 1 is a flowchart showing an example of an embodiment of a data compression method according to the present invention.

FIG. 2 is an explanatory diagram showing an outline of a data configuration of LSI layout data compressed by a data compression method according to the present invention.

FIGS. 3A and 3B are explanatory diagrams showing a specific example in a case where the data compression method according to the present invention is applied to a graphic in LSI layout data, where FIG. (C) is a diagram showing instance information in the work area, (d) is a diagram when applied to an Xtrap trapezoid, (e) is a diagram when applied to a Ytrap trapezoid, (f) ) Is a diagram when applied to a polygon with xytype = 0, (g) is a diagram when applied to a polygon with xytype = 1, (h) is a diagram when applied to a path with xytype = 0,
(I) is a diagram when applied to a path of xytype = 1, (j)
Is a diagram when applied to a path of xytype = 2, and (k) is a diagram when applied to cell quotation.

FIG. 4 is an explanatory diagram of an example in which LSI layout data after compression is dumped.

FIG. 5 is an explanatory diagram of an example illustrating the LSI layout data of FIG. 4;

FIG. 6 is a flowchart showing another embodiment of the data compression method according to the present invention.

FIG. 7 is an explanatory diagram showing a specific example of information in a work area when executing the data compression method shown in FIG. 6;
Is a diagram showing sequence information, (b) is a diagram (1) showing instance information, and (c) is a diagram (2) showing instance information.

FIG. 8 is an explanatory diagram showing an example of a rectangle whose arrangement is specified by LSI layout data.

[Explanation of symbols]

S106 Cell name extraction S107 Search and additional registration of cell name information S108 Register cell name INDEX in instance information S109 Shape information extraction S110 Search and additional registration of shape information S111 Register Offset and shape INDEX in instance information S112 Array information, rotation information Extraction S113 Search for sequence information and rotation information and additionally register S114 Register array INDEX in instance information and rotation INDEX in shape information

Claims (5)

[Claims] 1. A data compression method for data indicating a graphic arrangement in a predetermined area, wherein an index for identifying the shape of the graphic is provided for each of the shapes, and the index and the index of the graphic in the predetermined area are provided. A data compression method, wherein the data is expressed using start point position coordinates. 2. The method according to claim 1, wherein, when the index is given, a shape of a figure to which the index is given is defined by using a displacement between vertex coordinates in the shape.
Data compression method described. 3. The data compression according to claim 1, wherein, when the predetermined area is composed of a plurality of layers, different indexes are given to different layers arranged even if the figure has the same shape. Method. 4. The data compression according to claim 1, wherein an arrangement of cells as a set of a plurality of figures is regarded as one figure shape, and an index is given to each cell and to each rotation information thereof. Method. 5. When a plurality of figures having the same shape are arranged in a predetermined pattern in the predetermined area, an index is also given to the predetermined pattern, and the index is combined with the index given to the same shape. 2. The data compression method according to claim 1, wherein the data is represented by the data.