Semiconductor storage

1. Field of the Invention

The present invention relates to a semiconductor storage and more particularly to a combination of a plurality of memory cell structures of an MOS static RAM.

2. Description of the Background Art

For a semiconductor storage comprising plural kinds of memory cells having different port structures and the like, conventionally, a semiconductor storage has been disclosed in Patent Document 1 (Japanese Patent Application Laid-Open No. 6-349275 (1994) gazette), for example. The semiconductor storage comprises a 3-port cell section which is 3-port simultaneous accessible and a 1-port cell section having one access port, and can be implemented by connecting them to at least a pair of bit lines in common.

On the other hand, it has been demanded that data to be accessed on a bit length unit are divided on a several-bit unit and individual data are accessed on the several-bit unit thus obtained by the division in a computer field or the like. Moreover, it has also been demanded that a part of bits are constituted to be accessible from a plurality of ports.

In the semiconductor storage of the Patent Document 1 described above, however, the 3-port cell section and the 1-port cell section share a bit line. For this reason, there is a problem in that both of the port cell sections cannot be accessed at the same time and the demand cannot be met.

It is an object of the present invention to obtain a semiconductor storage having a one-chip structure in which memory cells present in different memory cell arrays can be accessed at the same time.

According to the present invention, a semiconductor storage includes first and second memory cell arrays and a first row decoder, and the first memory cell array includes a plurality of first memory cells and a plurality of first word lines. The plurality of first memory cells are arranged in at least one column over a plurality of rows and the plurality of first word lines are connected to the first memory cells on a row unit. On the other hand, the second memory cell array includes a plurality of second memory cells, a plurality of second and third word lines, and a plurality of first and second bit lines. The plurality of second memory cells are arranged in a matrix, the second word lines are connected to the second memory cells on a row unit, the plurality of third word lines are connected to the second memory cells on a row unit and are not connected to any of the plurality of first memory cells, and the plurality of first bit lines are provided corresponding to the second memory cells on a column unit and are accessible to the second memory cell connected to the second word line in a selection state out of the plurality of second word lines, and the plurality of second bit lines are provided corresponding to the second memory cells on a column unit and are accessible to the second memory cell connected to the third word line in a selection state out of the plurality of third word lines. The first row decoder simultaneously brings any of the first word lines and any of the second word lines into the selection state based on a first address signal.

The first and second memory cell arrays can be controlled by the first row decoder. Consequently, it is possible to obtain a semiconductor storage having a 1-chip structure which can be accessed to the first and second memory cells at the same time.

These and other objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.

FIG. 1 is a block diagram showing a structure of a semiconductor storage according to a first embodiment of the present invention,

FIG. 2 is a circuit diagram showing a memory cell array structure according to the first embodiment,

FIG. 3 is an explanatory view showing a layout structure provided under a first aluminum wiring layer in a 1-port memory cell as seen on a plane,

FIG. 4 is an explanatory view showing a layout structure provided above a second aluminum wiring layer in the 1-port memory cell,

FIG. 5 is an explanatory view showing a layout structure provided under a first aluminum wiring layer in a 2-port memory cell as seen on a plane,

FIG. 6 is an explanatory view showing a layout structure provided above the second aluminum wiring layer in the 2-port memory cell,

FIG. 7 is an explanatory view showing a layout structure provided under a first aluminum wiring layer in a 1-port and 2-port memory cell as seen on a plane,

FIG. 8 is a block diagram showing a structure of a semiconductor storage according to a second embodiment of the present invention,

FIG. 9 is a block diagram showing a structure of a semiconductor storage according to a third embodiment of the present invention,

FIG. 10 is an explanatory diagram showing an internal structure of a row decoder in FIG. 9 ,

FIG. 11 is a block diagram showing a structure of a semiconductor storage according to a fourth embodiment of the present invention,

FIG. 12 is a circuit diagram showing a memory cell array structure according to the fourth embodiment,

FIG. 13 is a circuit diagram showing the details of a CAM memory cell,

FIG. 14 is an explanatory view showing a layout structure of all layers in the CAM memory cell as seen on a plane,

FIG. 15 is an explanatory view showing a layout structure provided under a first aluminum wiring layer in FIG. 14 as seen on a plane,

FIG. 16 is an explanatory view showing a layout structure provided above a second aluminum wiring layer in FIG. 14 as seen on a plane,

FIG. 17 is a block diagram showing a structure of a semiconductor storage according to a fifth embodiment of the present invention,

FIG. 18 is a circuit diagram showing a memory cell array structure according to the fifth embodiment,

FIG. 19 is a block diagram showing a structure of a semiconductor storage according to a sixth embodiment of the present invention,

FIG. 20 is a circuit diagram showing a structure of a 2-port memory cell according to the sixth embodiment,

FIG. 21 is an explanatory view showing a layout structure of all layers of the 2-port memory cell according to the sixth embodiment as seen on a plane,

FIG. 22 is an explanatory view showing a layout structure provided under a first aluminum wiring layer in FIG. 21 as seen on a plane,

FIG. 23 is an explanatory view showing a layout structure provided above the first aluminum wiring layer in FIG. 21 as seen on a plane,

FIG. 24 is a block diagram showing a structure of a semiconductor storage according to a seventh embodiment of the present invention,

FIG. 25 is a circuit diagram showing a structure of a dummy cell formed in a memory cell area for timing regulation, and

FIG. 26 is a circuit diagram showing a memory cell array structure according to the seventh embodiment.

<First Embodiment>

(Whole Structure)

FIG. 1 is a block diagram showing a structure of a semiconductor storage according to a first embodiment of the present invention, As shown in FIG. 1 , a 1-port memory cell array 11 (a first memory cell array) and a 2-port memory cell array 12 (a second memory cell array) are provided together over one chip to constitute a memory macro. More specifically, the 1-port memory cell array 11 is provided with a word line WL 1 for a first port (a first word line (a common word line)) and the 2-port memory cell array 12 is provided with a word line WL 1 for a first port (a second word line (a common word line)) and a word line WL 2 for a second port (a third word line).

Upon receipt of an address input bus signal AD 1 , a control circuit 31 supplies a row address to a row decoder 16 (a first row decoder) and a column address to column selectors 21 and 22 under timing control of a read control input signal RE 1 and a write control input signal WE 1 .

Upon receipt of an address input bus signal AD 2 , a control circuit 32 supplies a row address to a row decoder 17 (a second row decoder) and a column address to a column selector 23 under timing control of a read control input signal RE 2 and a write control input signal WE 2 .

The row decoder 16 selectively sets any of a plurality of the word lines WL 1 for the first port into an active state (a selection state) based on the row address and the row decoder 17 selectively sets any of a plurality of the word lines WL 2 for the second port into the active state based on the row address.

The column selector 21 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 1 and bar BL 1 for the first port based on the column address sent from the control circuit 31 and inputs/outputs a data input/output bus signal DIO 11 < 7 : 4 > to/from the 1-port memory cell array 11 through the bit line pair BL 1 and bar BL 1 for the first port thus selected.

The column selector 22 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 21 and bar BL 21 for the first port (first bit lines) based on the column address sent from the control circuit 31 and inputs/outputs a data input/output bus signal DIO 12 < 3 : 0 > to/from the 2-port memory cell array 12 through the bit line pairs BL 21 and bar BL 21 for the first port thus selected.

The column selector 23 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 22 and bar BL 22 for the second port (second bit lines) based on the column address sent from the control circuit 32 and inputs/outputs a data input/output bus signal DIO 2 < 3 : 0 > to/from the 2-port memory cell array 12 through the bit line pairs BL 22 and bar BL 22 for the second port thus selected.

These column selectors 22 to 24 have a sense amplifier (SA) function and a write driving (WD) function.

(Memory Cell Structure)

FIG. 2 is a circuit diagram showing a memory cell array structure of the 1-port memory cell array 11 and the 2-port memory cell array 12 . For convenience of explanation, FIG. 2 shows only a 2×2 memory cell but does not imply an actual size of the memory cell array (a matrix-like memory cell arrangement).

The 1-port memory cell array 11 is constituted by memory cells ms 00 , ms 01 , ms 10 and ms 11 .

Each of the memory cells ms 00 to ms 11 is constituted by inverters 41 and 42 which are cross connected to each other, an NMOS transistor Q 11 having one of electrodes connected to an input of the inverter 41 (an output of the inverter 42 ), and an NMOS transistor Q 12 having one of electrodes connected to an input of the inverter 42 (an output of the inverter 41 ).

Word lines WL 1 < 0 > and WL 1 < 1 > for the first port are driven by drivers 52 and 51 (which are usually provided in the row decoder 16 (not shown)), respectively.

In the memory cells ms 00 and ms 01 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 < 0 > for the first port in common. In the memory cells ms 10 and ms 11 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 < 1 > for the first port in common.

Both of the memory cells ms 00 and ms 10 on the same column are provided between a bit line pair BL 1 < 0 > and bar BL 1 < 0 > for the first port. In these memory cells ms 00 and ms 10 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 1 < 0 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 1 < 0 > for the first port.

Both of the memory cells ms 01 and ms 11 on the same column are provided between a bit line pair BL 1 < 1 > and bar BL 1 < 1 > for the first port. In the memory cells ms 01 and ms 11 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 1 < 1 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 1 < 1 > for the first port.

On the other hand, the 2-port memory cell array 12 is constituted by memory cells md 00 , md 01 , md 10 and md 11 .

Each of the memory cells md 00 to md 11 is constituted by the inverters 41 and 42 which are cross connected to each other, NMOS transistors Q 11 and Q 21 having electrodes connected to the input of the inverter 41 , and NMOS transistors Q 12 and Q 22 having electrodes connected to the input of the inverter 42 .

Word lines WL 2 < 0 > and WL 2 < 1 > for the second port are driven by drivers 54 and 53 (which are usually provided in the row decoder 17 (not shown)), respectively.

In the memory cells md 00 and md 01 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 < 0 > for the first port in common, and gate electrodes of the NMOS transistors Q 21 and Q 22 are electrically connected to the word line WL 2 < 0 > for the second port in common.

In the memory cells md 10 and md 11 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 < 1 > for the first port in common, and gate electrodes of the NMOS transistors Q 21 and Q 22 are electrically connected to the word line WL 2 < 1 > for the second port in common.

Both of the memory cells md 00 and md 10 on the same column are provided between a bit line pair BL 21 < 0 > and bar BL 21 < 0 > for the first port. In these memory cells md 00 and md 10 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 21 < 0 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 21 < 0 > for the first port.

Furthermore, both of the memory cells md 00 and md 10 are provided between a bit line pair BL 22 < 0 > and bar BL 22 < 0 > for the second port. In these memory cells md 00 and md 10 , the other electrode of the NMOS transistor Q 21 is electrically connected to the bit line BL 22 < 0 > for the second port and the other electrode of the NMOS transistor Q 22 is electrically connected to the inverted bit line bar BL 22 < 0 > for the second port.

Both of the memory cells md 01 and md 11 on the same column are provided between a bit line pair BL 21 < 1 > and bar BL 21 < 1 > for the first port. In these memory cells md 01 and md 11 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 21 < 1 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 21 < 1 > for the first port.

Furthermore, both of the memory cells md 01 and md 11 are provided between a bit line pair BL 22 < 1 > and bar BL 22 < 1 > for the second port. In these memory cells md 10 and md 11 , the other electrode of the NMOS transistor Q 21 is electrically connected to the bit line BL 22 < 1 > for the second port and the other electrode of the NMOS transistor Q 22 is electrically connected to the inverted bit line bar BL 22 < 1 > for the second port.

In FIGS. 1 and 2 , and FIG. 3 and subsequent drawings, a symbol < > represents a bus signal (line) and a numeral in < > corresponds to a bus signal name. Moreover, <n−1:0> represents a signal having an n-bit width of bus signals 0 to (n−1).

(Operation)

With reference to FIGS. 1 and 2 , description will be given to a reading and writing operation using the first port and a reading and writing operation using the second port. For convenience of explanation, a memory cell array structure of the 1-port memory cell array 11 and the 2-port memory cell array 12 is set to be n (rows)×m (columns).

First of all, in the case in which the reading operation of the first port is to be executed, the read control input signal RE 1 is enabled. When the read control input signal RE 1 is brought into the enable state, the row decoder 16 for receiving a row address corresponding to the address input bus signal AD 1 through the control circuit 31 selectively sets, into an active state, a word line WL 1 <i> (i=0 to n−1) for the first port in word lines WL 1 <n−1:0> for the first port having n bits.

Consequently, the NMOS transistors Q 11 and Q 12 of each of selected memory cells msi 0 to msim of the 1-port memory cell array 11 connected to the word line WL 1 <i> for the first port which is set into the active state and selected memory cells mdi 0 to mdim of the 2-port memory cell array 12 are turned ON, and contents stored in each of the selected memory cells are propagated to corresponding bit line pairs BL 1 and bar BL 1 , and BL 21 and bar BL 21 for the first port.

Each of the column selectors 21 and 22 for receiving the column address corresponding to the address input bus signal AD 1 through the control circuit 31 selects four sets of bit line pairs from m bit line pairs BL 1 and bar BL 1 , and BL 21 and bar BL 21 for the first port, and senses and amplifies them by means of an internal sense amplifier. Consequently, the data input/output bus signal DIO 11 < 7 : 4 > and the data input/output bus signal DIO 12 < 3 : 0 > are output as 8-bit read data.

In the case in which the writing operation of the first port is to be executed, next, the write control input signal WE 1 is enabled. When the write control input signal WE 1 is brought into the enable state, the row decoder 16 for receiving a row address corresponding to the address input bus signal AD 1 through the control circuit 31 selectively sets, into an active state, the word line WL 1 <i> (i=0 to n−1) for the first port in the word lines WL 1 <n−1:0> for the first port having n bits.

Consequently, the NMOS transistors Q 11 and Q 12 of each of the selected memory cells msi 0 to msim of the 1-port memory cell array 11 and the selected memory cells mdi 0 to mdim of the 2-port memory cell array 12 which are connected to the word line WL 1 <i> for the first port set into the active state are turned ON, and the inverters 41 and 42 of each selected memory cell are electrically connected to the bit line pairs BL 1 and bar BL 1 , and BL 21 and bar BL 21 for the first port.

At the same time, write data sent from the outside are given as the data input/output bus signal DIO 11 < 7 : 4 > and the data input/output bus signal DIO 12 < 3 : 0 > to/from the column selectors 21 and 22 .

Each of the column selectors 21 and 22 for receiving the column address corresponding to the address input bus signal AD 1 through the control circuit 31 selects four sets of bit line pairs from m bit line pairs BL 1 and bar BL 1 , and BL 21 and bar BL 21 for the first port, and gives the write data to the selected bit line pairs, thereby driving the selected bit line pairs to “L” or “H”. As a result, it is possible to carry out writing based on the write data for the selected memory cell.

Then, in the case in which the reading operation of the second port is to be executed, the read control input signal RE 2 is enabled. When the read control input signal RE 2 is brought into the enable state, the row decoder 17 for receiving a row address corresponding to the address input bus signal AD 2 through the control circuit 32 selectively sets, into an active state, a word line WL 2 <i> (i=0 to n−1) for the second port in word lines WL 2 <n−1: 0 > for the second port having n bits.

Consequently, the NMOS transistors Q 21 and Q 22 of each of selected memory cells mdi 0 to mdim of the 2-port memory cell array 12 connected to the word line WL 2 <i> for the second port which is set into the active state are turned ON, and contents stored in each of the selected memory cells are propagated to corresponding bit line pairs BL 22 and bar BL 22 for the second port.

Subsequently, the column selector 23 for receiving a column address corresponding to the address input bus signal AD 2 through the control circuit 32 selects four sets of bit line pairs from m bit line pairs BL 22 and bar BL 22 for the second port respectively, and senses and amplifies them by means of an internal sense amplifier. Consequently, the data input/output bus signal DIO 2 < 3 : 0 > are output as 4-bit read data.

In the case in which the writing operation of the second port is to be executed, next, the write control input signal WE 2 is enabled. When the write control input signal WE 2 is brought into the enable state, the row decoder 17 for receiving a row address corresponding to the address input bus signal AD 2 through the control circuit 32 selectively sets, into an active state, the word line WL 2 <i> for the second port in the word lines WL 2 <n−1: 0 > for the second port having n bits.

Consequently, the NMOS transistors Q 21 and Q 22 of each of the selected memory cells mdi 0 to mdim of the 2-port memory cell array 12 connected to the word line WL 2 <i> for the second port which is set into the active state are turned ON, and the inverters 41 and 42 of each selected memory cell are electrically connected to the bit line pair BL 22 and bar BL 22 for the second port.

At the same time, write data sent from the outside are given as the data input/output bus signal DIO 2 < 3 : 0 > to/from the column selector 23 .

The column selector 23 for receiving a column address corresponding to the address input bus signal AD 2 through the control circuit 32 selects four sets of bit line pairs from m bit line pairs BL 22 and bar BL 22 for the second port, and gives the write data to the selected bit line pairs, thereby driving the selected bit line pairs to “L” or “H”. As a result, it is possible to carry out writing based on the write data for a selected memory cell.

As described above, it is possible to read and write 8-bit data which are divided and held in high and low orders on a 4-bit unit in the 1-port memory cell array 11 and the 2-port memory cell array 12 from the first port.

Then, it is possible to read and write the lower 4-bit data which are held in the 2-port memory cell array 12 from the second port. In this case, the upper 4-bit data are not accessed at all.

As a result, the semiconductor storage according to the first embodiment can divide, on a 4-bit unit, data to be accessed on an 8-bit length unit and can access data on a lower 4-bit unit thus obtained by the division. The lower 4-bit is accessible from the first and second ports so that it is possible to carry out reading and writing to meet a demand for a data dividing access.

In the first embodiment, thus, the 1-port memory cell array 11 and the 2-port memory cell array 12 including memory cells having different circuit structures from each other can be controlled simultaneously by the row decoder 16 . Therefore, it is possible to obtain a semiconductor storage having a 1-chip structure in which a single port memory cell and a 2-port memory cell having different structures can be accessed simultaneously.

Even if the 1-port memory cell array 11 is constituted with a replacement by a 2-port memory cell, moreover, an equivalent operation to that of the first embodiment can be carried out. As compared with this case, an area can be decreased because the upper 4 bits are constituted by the 1-port memory cell, and a layout area can be decreased because it is possible to eliminate a control circuit, a column selector and the like for the upper 4 bits which do not need to be accessed in the second port. Since a circuit scale is thus decreased, it is possible to reduce power consumption during an operation and a standby.

(Layout Structure)

FIGS. 3 and 4 are explanatory views showing a layout structure of a single port memory cell MS of the 1-port memory cell array 11 . FIG. 3 is an explanatory view mainly showing a layout structure provided under a first aluminum wiring layer as seen on a plane. FIG. 4 is an explanatory view showing a layout structure provided above a second aluminum wiring layer as seen on a plane.

It is assumed that the inverter 41 shown in FIG. 2 is a CMOS inverter constituted by a PMOS transistor P 1 and an NMOS transistor N 1 and the inverter 42 is a CMOS inverter constituted by a PMOS transistor P 2 and an NMOS transistor N 2 .

As shown in FIG. 3 , the PMOS transistors P 1 and P 2 are formed in an N well region NW, the NMOS transistor N 1 and the NMOS transistor Q 11 are formed in a P well region PW 1 , and the NMOS transistor N 2 and the NMOS transistor Q 12 are formed in a P well region PW 0 . The P well region PW 0 and the P well region PW 1 are formed opposite to each other with the N well region NW interposed therebetween.

In the N well region NW, the PMOS transistor P 1 is constituted by a polysilicon wiring PL 1 provided across a P + diffusion region FLP 1 and the PMOS transistor P 2 is constituted by a polysilicon wiring PL 2 provided across a P + diffusion region FLP 2 .

In the P well region PW 1 , the NMOS transistor N 1 is constituted by the polysilicon wiring PL 1 provided across an N + diffusion region FLN 1 and the NMOS transistor Q 11 is constituted by a polysilicon wiring PL 11 provided across an N + diffusion region FLQ 11 . The polysilicon wiring PL 1 is formed from the N well region NW to the P well region PW 1 and is thus shared as gates of the NMOS transistor N 1 and the PMOS transistor P 1 .

In the P well region PW 0 , the NMOS transistor N 2 is constituted by the polysilicon wiring PL 2 provided across an N + diffusion region FLN 2 and the NMOS transistor Q 12 is constituted by a polysilicon wiring PL 12 provided across an N + diffusion region FLQ 12 . The polysilicon wiring PL 2 is formed from the N well region NW to the P well region PW 0 and is thus shared as gates of the NMOS transistor N 2 and the PMOS transistor P 2 .

Moreover, the polysilicon wiring PL 1 and the polysilicon wiring PL 12 are formed on the same straight line, the polysilicon wiring PL 2 and the polysilicon wiring PL 11 are formed on the same straight line, the diffusion regions FLP 1 , FLN 1 and FLQ 12 are formed on the same straight line in almost the same form, and the diffusion regions FLP 2 , FLN 2 and FLQ 11 are formed on the same straight line in almost the same form.

As a result, the PMOS transistor P 1 , the NMOS transistor N 1 and the NMOS transistor Q 12 can be formed along the same straight line and the PMOS transistor P 2 , the NMOS transistor N 2 and the NMOS transistor Q 11 can be formed along the same straight line. Consequently, a cell height HC 1 of the single port memory cell MS can be set corresponding to two transistors. In this specification, the cell height implies a formation length in a direction of formation of a bit line (a longitudinal direction in the drawing) on the layout structure.

The P + diffusion regions FLP 1 and FLP 2 are obtained by implanting and diffusing a P-type impurity and the N + diffusion regions FLN 1 , FLN 2 , FLQ 11 and FLQ 12 are obtained by implanting and diffusing an N-type impurity. In the description of FIG. 3 , each diffusion region has an upper region referred to as one of regions and a lower region referred to as the other region with respect to the polysilicon wiring in FIG. 3 .

In the P well region PW 1 , a ground wiring LG 1 (a first layer aluminum wiring) provided on one of the regions of the N + diffusion region FLN 1 is electrically connected to the diffusion region FLN 1 through a diffusion contact hole CH. The polysilicon wiring PL 11 is electrically connected to a word line 1 WL 1 (the first layer aluminum wiring) through a gate contact hole GC, and a bit line 1 BL 1 (the first layer aluminum wiring) provided on the other region of the diffusion region FLQ 11 is electrically connected to the other region of the diffusion region FLQ 11 through the diffusion contact hole CH.

The diffusion contract hole CH implies a contact hole of the diffusion region and the first layer (aluminum) wiring, and the gate contact hole GC implies a contact hole of the polysilicon wiring and the first layer wiring.

In the N well region NW, a power wiring LV 1 (a first layer aluminum wiring) provided on one of the regions of the diffusion region FLP 1 is electrically connected to one of the regions of the diffusion region FLP 1 through a diffusion contact hole CH, and the power wiring LV 1 provided on the other region of the diffusion region FLP 2 is electrically connected to the other region of the diffusion region FLP 2 through the diffusion contact hole CH.

In the P well region PW 0 , the inverted bit line bar BL 1 (a first layer aluminum wiring) provided on one of the regions of the diffusion region FLQ 12 is electrically connected to one of the regions of the diffusion region FLQ 1 through a diffusion contact hole CH, and the word line 1 WL 1 provided on the polysilicon wiring PL 12 is electrically connected to the polysilicon wiring PL 12 through a gate contact hole GC. A ground wiring LG 1 provided on the other region of the diffusion region FLN 2 is electrically connected to the other region of the diffusion region FLN 2 through a contact hole CH.

The aluminum wiring AL 11 to be the first layer aluminum wiring extended from the other region of the diffusion region FLN 1 to the other region of the P + diffusion region FLP 1 is electrically connected to one of the regions of the diffusion region FLN 1 through the diffusion contact hole CH. The aluminum wiring AL 11 is electrically connected to the other region of the P + diffusion region FLP 1 and the polysilicon wiring PL 2 through a shared contact SC formed from the other region of the P + diffusion region FLP 1 to the polysilicon wiring PL 2 .

The aluminum wiring AL 12 to be the first layer aluminum wiring extended from one of the regions of the diffusion region FLN 2 to one of the regions of the P + diffusion region FLP 2 is electrically connected to one of the regions of the diffusion region FLN 2 through the diffusion contact hole CH. The aluminum wiring AL 12 is electrically connected to one of the regions of the P + diffusion region FLP 2 and the polysilicon wiring PL 1 through the shared contact SC formed from one of the regions of the P + diffusion region FLP 2 to the polysilicon, wiring PL 1 .

As shown in FIG. 4 , a ground wiring LG 2 (a second layer aluminum wiring) is electrically connected to the lower ground wiring LG 1 (not shown) through a via hole VH 1 . A power wiring LV 2 (the second layer aluminum wiring) is electrically connected to the lower power wiring LV 1 (not shown) through the via hole VH 1 .

A word line 2 WL 1 (the second layer aluminum wiring) is electrically connected to the word line 1 WL 1 (not shown) through the via hole VH 1 , and a word line 3 WL 1 (a third layer aluminum wiring) is electrically connected to the word line 2 WL 1 through a via hole VH 2 . These word lines 1 WL 1 to 3 WL 1 constitute the word line WL 1 for the first port in FIGS. 1 and 2 .

The word line 3 WL 1 is formed across the P well regions PW 0 and PW 1 and the N well region NW. The via hole VH 1 implies a via hole for connecting the first layer wiring to the second layer (aluminum) wiring, and the via hole VH 2 implies a via hole for connecting the second layer wiring to the third layer (aluminum) wiring.

The bit line 2 BL 1 (the second layer aluminum wiring) is electrically connected to the lower bit line 1 BL 1 (not shown) through the via hole VH 1 and the inverted bit line bar 2 BL 1 (the second layer aluminum wiring) is electrically connected to the lower inverted bit line bar 1 BL 1 (not shown) through the via hole VH 1 .

The bit line pair BL 1 and bar BL 1 for the first port in FIGS. 1 and 2 are constituted by the bit line 2 BL 1 , the bit line 1 BL 1 and the inverted bit lines bar 2 BL 1 and bar 1 BL 1 .

The bit lines 2 BL 1 and bar 2 BL 1 , the ground wiring LG 2 and the power wiring LV 2 are formed in parallel with each other in a longitudinal direction of FIG. 4 over the P well regions PW 1 and PW 0 and the N well region NW, respectively.

FIGS. 5 and 6 are explanatory views showing a layout structure of a 2-port memory cell MD of the 2-port memory cell array 12 . FIG. 5 is an explanatory view mainly showing a layout structure provided under the first aluminum wiring layer as seen on a plane. FIG. 6 is an explanatory view showing a layout structure provided above the second aluminum wiring layer as seen on a plane.

It is assumed that the inverter 41 shown in FIG. 2 is a CMOS inverter constituted by a PMOS transistor P 1 and an NMOS transistor N 1 and the inverter 42 is a CMOS inverter constituted by a PMOS transistor P 2 and an NMOS transistor N 2 .

As shown in FIG. 5 , the PMOS transistors P 1 and P 2 are formed in the N well region NW, the NMOS transistor N 1 and the NMOS transistors Q 21 and Q 22 are formed in the P well region PW 1 , and the NMOS transistor N 2 and the NMOS transistors Q 11 and Q 12 are formed in the P well region PW 0 . The P well region PW 0 and the P well region PW 1 are formed opposite to each other with the N well region NW interposed therebetween.

In the N well region NW, the PMOS transistor P 1 is constituted by a polysilicon wiring PL 1 provided across a P + diffusion region FLP 1 and the PMOS transistor P 2 is constituted by a polysilicon wiring PL 2 provided across a P + diffusion region FLP 2 .

In the P well region PW 0 , the NMOS transistor N 1 is constituted by the polysilicon wiring PL 1 provided across an N + diffusion region FLN 1 and the NMOS transistors Q 11 and Q 12 are constituted by a polysilicon wiring PL 10 provided across N + diffusion regions FLQ 11 and FLQ 12 . The polysilicon wiring PL 1 is formed from the N well region NW to the P well region PW 1 and is thus shared as gates of the NMOS transistor N 1 and the PMOS transistor P 1 , and the polysilicon wiring PL 10 is shared as gates of the NMOS transistors Q 11 and Q 12 .

In the P well region PW 1 , the NMOS transistor N 2 is constituted by the polysilicon wiring PL 2 provided across an N + diffusion region FLN 2 and the NMOS transistors Q 21 and Q 22 are constituted by a polysilicon wiring PL 20 provided across N + diffusion regions FLQ 21 and FLQ 22 , respectively. The polysilicon wiring PL 2 is formed from the N well region NW to the P well region PW 0 and is thus shared as gates of the NMOS transistor N 2 and the PMOS transistor P 2 , and the polysilicon wiring PL 20 is shared as gates of the NMOS transistors Q 21 and Q 22 .

Moreover, the polysilicon wiring PL 1 and the polysilicon wiring PL 20 are formed on the same straight line, the polysilicon wiring PL 2 and the polysilicon wiring PL 10 are formed on the same straight line, the diffusion regions FLP 1 , FLN 1 , FLQ 21 and FLQ 22 are formed on the same straight line, and the diffusion regions FLP 2 , FLN 2 , FLQ 11 and FLQ 12 are formed on the same straight line.

As a result, the PMOS transistor P 1 and the NMOS transistors N 1 , Q 21 and Q 22 can be formed along the same straight line and the PMOS transistor P 2 and the NMOS transistors N 2 , Q 11 and Q 12 can be formed along the same straight line. Consequently, a cell height HC 2 of the 2-port memory cell MD can be set corresponding to two transistors in the same manner as the single port memory cell MS (HC 2 =HC 1 ).

The P + diffusion regions FLP 1 and FLP 2 are obtained by implanting and diffusing a P-type impurity and the N + diffusion regions FLN 1 , FLN 2 , FLQ 11 , FLQ 12 , FLQ 21 and FLQ 22 are obtained by implanting and diffusing an N-type impurity. In the description of FIG. 5 , each diffusion region has an upper region referred to as one of regions and a lower region referred to as the other region with respect to the polysilicon wiring in FIG. 5 .

In the P well region PW 0 , a ground wiring LG 1 provided on one of the regions of the N + diffusion region FLN 1 is electrically connected to one of the regions of the diffusion region FLN 1 through a diffusion contact hole CH. The polysilicon wiring PL 10 is electrically connected to a word line 1 WL 1 through a gate contact hole GC, and a bit line 1 BL 21 (the first layer aluminum wiring) provided on the other region of the diffusion region FLQ 11 is electrically connected to the other region of the diffusion region FLQ 11 through the diffusion contact hole CH, and the inverted bit line bar 1 BL 21 (the first layer aluminum wiring) provided on the other region of the diffusion region FLQ 12 is electrically connected to the other region of the diffusion region FLQ 12 through the diffusion contact hole CH.

Furthermore, the polysilicon wiring PL 1 and one of the regions of the N + diffusion region FLQ 12 are electrically connected to each other through the shared contact SC formed from one of the regions of the N + diffusion region FLQ 11 to the polysilicon wiring PL 1 .

In the N well region NW, a power wiring LV 1 provided on one of the regions of the diffusion region FLP 1 is electrically connected to one of the regions of the diffusion region FLP 1 through the diffusion contact hole CH, and the power wiring LV 1 provided on the other region of the diffusion region FLP 2 is electrically connected to the other region of the diffusion region FLP 2 through the diffusion contact hole CH.

In the P well region PW 1 , the bit line 1 BL 22 (the first layer aluminum wiring) provided on one of the regions of the diffusion region FLQ 21 is electrically connected to one of the regions of the diffusion region FLQ 21 through the diffusion contact hole CH, and the inverted bit line bar 1 BL 22 (the first layer aluminum wiring) provided on one of the regions of the diffusion region FLQ 22 is electrically connected to one of the regions of the diffusion region FLQ 22 through the diffusion contact hole CH.

The word line 1 WL 2 (the first layer aluminum wiring) provided on the polysilicon wiring PL 20 is electrically connected to the polysilicon wiring PL 20 through the gate contact hole GC, and the ground wiring LG 1 provided on the other region of the diffusion region FLN 2 is electrically connected to the other region of the diffusion region FLN 2 through the diffusion contact hole CH.

The aluminum wiring AL 11 to be the first layer aluminum wiring extended from the other region of the diffusion region FLN 1 to the other region of the P + diffusion region FLP 1 is electrically connected to one of the regions of the diffusion region FLN 1 through the diffusion contact hole CH. The aluminum wiring AL 11 is electrically connected to the other region of the P + diffusion region FLP 1 and the polysilicon wiring PL 2 through a shared contact SC formed from the other region of the P + diffusion region FLP 1 to the polysilicon wiring PL 2 .

The aluminum wiring AL 12 to be the first layer aluminum wiring extended from one of the regions of the diffusion region FLN 2 to one of the regions of the P + diffusion region FLP 2 is electrically connected to one of the regions of the diffusion region FLN 2 through the diffusion contact hole CH. The aluminum wiring AL 12 is electrically connected to one of the regions of the P + diffusion region FLP 2 and the polysilicon wiring PL 1 through the shared contact SC formed from one of the regions of the P + diffusion region FLP 2 to the polysilicon wiring PL 1 .

As shown in FIG. 6 , a ground wiring LG 2 is electrically connected to the lower ground wiring LG 1 (not shown) through a via hole VH 1 . A power wiring LV 2 is electrically connected to the lower power wiring LV 1 (not shown) through the via hole VH 1 .

A word line 2 WL 1 is electrically connected to the word line 1 WL 1 (not shown) through the via hole VH 1 , and a word line 3 WL 1 is electrically connected to the word line 2 WL 1 through a via hole VH 2 . These word lines 1 WL 1 to 3 WL 1 constitute the word line WL 1 for the first port in FIGS. 1 and 2 .

A word line 2 WL 2 (the second layer aluminum wiring) is electrically connected to the word line 1 WL 2 (not shown) through the via hole VH 1 , and a word line 3 WL 2 (a third layer aluminum wiring) is electrically connected to the word line 2 WL 2 through the via hole VH 2 . These word lines 1 WL 2 to 3 WL 2 constitute the word line WL 2 for the second port in FIGS. 1 and 2 .

The word lines 3 WL 1 and 3 WL 2 are formed across the P well regions PW 0 and PW 1 and the N well region NW.

The bit line 2 BL 21 (the second layer aluminum wiring) is electrically connected to the lower bit line 1 BL 21 (not shown) through the via hole VH 1 and the inverted bit line bar 2 BL 21 (the second layer aluminum wiring) is electrically connected to the lower inverted bit line bar 1 BL 21 (not shown) through the via hole VH 1 .

The bit line 2 BL 22 (the second layer aluminum wiring) is electrically connected to the lower bit line 1 BL 22 (not shown) through the via hole VH 1 and the inverted bit line bar 2 BL 22 (the second layer aluminum wiring) is electrically connected to the lower inverted bit line bar 1 BL 22 (not shown) through the via hole VH 1 .

The bit line pair BL 21 and bar BL 21 for the first port in FIGS. 1 and 2 are constituted by the bit line 2 BL 21 , the bit line 1 BL 21 and the inverted bit lines bar 2 BL 21 and bar 1 BL 21 , and the bit line pair BL 22 and bar BL 22 for the second port in FIGS. 1 and 2 are constituted by the bit line 2 BL 22 , the bit line 1 BL 22 and the inverted bit lines bar 2 BL 22 and bar 1 BL 22 .

The bit lines 2 BL 21 and bar 2 BL 21 , the bit lines 2 BL 22 and bar 2 BL 22 , and the ground wiring LG 2 and the power wiring LV 2 are formed in parallel with each other in a longitudinal direction of the drawing over the P well regions PW 1 and PW 0 and the N well region NW, respectively.

FIG. 7 is an explanatory view showing a layout structure of a single port memory cell MS and a 2-port memory cell MD in the vicinity of a boundary between the 1-port memory cell array 11 and the 2-port memory cell array 12 . FIG. 7 is an explanatory view mainly showing a layout structure provided under the first aluminum wiring layer as seen on a plane.

As shown in FIG. 7 , cell heights of the single port memory cell MS and the 2 port memory cell MD are set to be equal to each other (HC 1 =HC 2 ) so that the single port memory cell MS and the 2-port memory cell MD can be arranged adjacently without a useless region such as a spacer. In this case, the word line 1 WL 1 , the gate contact hole GC and the via hole VH 1 (not shown) provided on a boundary line 39 are shared as shown in FIG. 7 .

<Second Embodiment>

FIG. 8 is a block diagram showing a structure of a semiconductor storage according to a second embodiment of the present invention. For the semiconductor storage according to the second embodiment, a 1-chip semiconductor storage is implemented by using a memory macro 60 (a first memory cell array) having a 1-port memory cell structure and a memory macro 70 (a second memory cell array) having a 2-port memory cell structure.

As shown in FIG. 8 , the semiconductor storage is implemented by a combination of the memory macros 60 and 70 provided independently. The memory macro 60 is constituted by a 1-port memory cell array 61 , a row decoder 62 , a control circuit 63 and a column selector 64 . The 1-port memory cell array 61 is provided With a word line WL 1 for a first port (a first word line), and the 2-port memory cell array 71 is provided with a word line WL 21 for the first port (a second word line) and a word line WL 22 for the second port (a third word line).

Upon receipt of an address input bus signal AD 1 , a control circuit 63 supplies a row address to a row decoder 62 and a column address to a column selector 64 under timing control of a read control input signal RE 1 and a write control input signal WE 1 . The row decoder 62 (a first row decoder (No. 1 )) selectively sets any of the plurality of the word lines WL 1 for the first port into an active state based on the row address.

The column selector 64 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 1 and bar BL 1 for the first port based on the column address sent from the control circuit 63 and inputs/outputs a data input/output bus signal DIO 11 < 7 : 4 > to/from the 1-port memory cell array 61 through the bit line pair BL 1 and bar BL 1 for the first port thus selected.

On the other hand, the memory macro 70 is constituted by a 2-port memory cell array 71 , a peripheral circuit for the first port (a row decoder 72 (a first row decoder (No. 2 )), a control circuit 73 and a column selector 74 ), and a peripheral circuit for the second port (a row decoder 75 (a second row decoder), a control circuit 76 and a column selector 77 ).

Upon receipt of an address input bus signal AD 1 , a control circuit 73 supplies a row address to a row decoder 72 and a column address to a column selector 74 under timing control of the read control input signal RE 1 and the write control input signal WE 1 . The row decoder 72 selectively sets any of a plurality of the word lines WL 21 for the first port into an active state based on the row address.

The column selector 74 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 21 and bar BL 21 for the first port based on the column address sent from the control circuit 73 and inputs/outputs a data input/output bus signal DIO 11 < 3 : 0 > to/from the 2-port memory cell array 71 through the bit line pairs BL 21 and bar BL 21 for the first port thus selected.

Upon receipt of an address input bus signal AD 2 , a control circuit 76 supplies a row address to a row decoder 75 and a column address to a column selector 77 under timing control of a read control input signal RE 2 and a write control input signal WE 2 . The row decoder 75 selectively sets any of a plurality of the word lines WL 22 for the second port into an active state based on the row address.

The column selector 77 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 22 and bar BL 22 for the second port based on the column address sent from the control circuit 76 and inputs/outputs a data input/output bus signal DIO 2 < 3 : 0 > to/from the 2-port memory cell array 71 through the bit line pairs BL 22 and bar BL 22 for the second port thus selected.

These column selectors 64 , 74 and 77 have a sense amplifier function and a write driving function.

An operation of the semiconductor storage according to the second embodiment is carried out basically in the same manner as that of the semiconductor storage according to the first embodiment except that simultaneous control of the 1-port memory cell array 11 and the 2-port memory cell array 12 by the row decoder 16 according to the first embodiment is replaced with simultaneous control of the 1-port memory cell array 61 and the 2-port memory cell array 71 by the two row decoders 62 and 72 .

In the semiconductor storage according to the second embodiment, accordingly, it is possible to read and write 8-bit data which are divided and held in high and low orders on a 4-bit unit in the 1-port memory cell array 61 and the 2-port memory cell array 71 from the first port based on the address input bus signal AD 1 .

Then, it is possible to read and write lower 4-bit data which are held in the 2-port memory cell array 71 from the second port based on the address input bus signal AD 2 . In this case, the upper 4-bit data are not accessed at all.

As a result, the semiconductor storage according to the second embodiment can divide, on a 4-bit unit, data to be accessed on an 8-bit length unit and can access data on a lower 4-bit unit thus obtained by the division. The lower 4 bits are accessible from the first and second ports so that it is possible to carry out reading and writing to meet a demand for a data dividing access.

In the second embodiment, thus, the memory macro 60 and the memory macro 70 are combined and the 1-port memory cell array 61 and the 2-port memory cell array 71 which have different port structures can be simultaneously controlled by the row decoders 62 and 63 based on the same address input bus signal AD 1 , respectively. Consequently, it is possible to obtain a semiconductor storage having a 1-chip structure in which the 1-port memory cell array 61 and the 2-port memory cell array 71 can be accessed at the same time.

As compared with the semiconductor storage according to the first embodiment, the semiconductor storage according to the second embodiment can produce an effect that the word lines WL 1 and WL 21 for the first port of the 1-port memory cell array 61 and the 2-port memory cell array 71 are individually driven by the row decoders 62 and 72 so that a signal propagation delay can be prevented from being increased due to a resistance component of the word line.

Furthermore, there can be produced an effect that a distance from the row decoders 62 and 72 according to the second embodiment to memory cells provided in the most distant positions of the 1-port memory cell array 61 and the 2-port memory cell array 71 can be set to be shorter than a distance from the row decoder 16 according to the first embodiment to a memory cell provided in the most distant position, resulting in prevention of an increase in the signal propagation delay due to the resistance component of the word line.

On the other hand, in the structure according to the first embodiment shown in FIG. 1 , the number of the row decoders and the control circuits can be reduced by one as compared with the structure shown in FIG. 8 (both of the numbers is two in FIG. 1 and both of the numbers is three in FIG. 8 ). Thus, the degree of integration can be more enhanced.

<Third Embodiment>

FIG. 9 is a block diagram showing a structure of a semiconductor storage according to a third embodiment of the present invention.

As shown in FIG. 9 , a 1-port memory cell array 11 L (a first memory cell array) and a 2-port memory cell array 12 R (a second memory cell array) are provided together over one chip with a row decoder 18 (a first row decoder) interposed therebetween, thereby constituting a memory macro. The 1-port memory cell array 11 L is provided with a word line WL 1 L for a first port (a first word line) and the 2-port memory cell array 12 R is provided with a word line WL 1 R for the first port (a second word line) and a word line WL 2 for the second port (a third word line).

The row decoder 18 drives the word lines WL 1 L and WL 1 R for the first port in such a manner that the same row is set into an active state in common. Other structures are the same as those in the first embodiment shown in FIG. 1 .

In the third embodiment, thus, the control for setting, into an active state, the word lines WL 1 L and WL 1 R for the first port which are provided in the 1-port memory cell array 11 L and the 2-port memory cell array 12 R is carried out in common on a row unit by the row decoder 18 arranged between the 1-port memory cell array 11 L and the 2-port memory cell array 12 R.

FIG. 10 is an explanatory diagram showing an internal structure of the row decoder 18 in FIG. 9 . As shown in FIG. 10 , the row decoder 18 has a decode circuit 18 a and selectively sets, into an “H” level, any of output lines OL< 0 > to OL<n−1> sent from a final stage logic gate group 65 based on an address input bus signal AD 1 obtained through a control circuit 30 .

The output lines OL< 0 > to OL<n−1> are connected to input sections of drivers DRL< 0 > to DRL<n−1> and to input sections of drivers DRR< 0 > to DRR<n−1>. The drivers DRL< 0 > to DRL<n−1> drive WL 1 L< 0 > to WL 1 L<n−1>, and the drivers DRR< 0 > to DRR<n−1> drive WL 1 R< 0 > to WL 1 R<n−1>. Accordingly, word lines WL 1 L<i> and WL 1 R<i> for the first port on the same row are selected (to have the “H” level) in common through a common output line OL<i>.

The semiconductor storage according to the third embodiment having such a structure further produces the following effects in addition to the effects of the first embodiment.

As compared with the semiconductor storage according to the first embodiment, the semiconductor storage according to the third embodiment can produce a first effect that the word lines WL 1 L and WL 1 R for the first port of the 1-port memory cell array 11 L and the 2-port memory cell array 12 R are driven individually and a signal propagation delay can be correspondingly prevented from being increased due to a resistance component of the word line.

Furthermore, it is possible to produce a second effect that a distance from the row decoder 18 according to the third embodiment to a memory cell provided in the most distant position can be set to be shorter than a distance from the row decoder 16 according to the first embodiment to a memory cell provided in the most distant position, resulting in prevention of an increase in the signal propagation delay due to the resistance component of the word line.

In the third embodiment, the separate word lines WL 1 L and WL 1 R for the first port have been provided for the 1-port memory cell array 11 L and the 2-port memory cell array 12 R. Also with such a structure that one word line WL 1 for the first port is provided on a row unit as a common word line to be shared by the 1-port memory cell array 11 L and the 2-port memory cell array 12 R and is driven by a common driver in the same manner as in the first embodiment, it is possible to produce the second effect that the signal propagation delay can be prevented from being increased due to the resistance component of the word line.

<Fourth Embodiment>

(Whole Structure)

FIG. 11 is a block diagram showing a structure of a semiconductor storage according to a fourth embodiment of the present invention.

As shown in FIG. 11 , a 1-port memory cell array 11 and a CAM (Content Addressable Memory) cell array 13 (a second memory cell array) are provided together over one chip to constitute a memory macro. More specifically, the 1-port memory cell array 11 is provided with a word line WL 1 for a first port (a first word line (a common word line)) and the CAM cell array 13 is provided with a word line WL 1 for a first port (a second word line (a common word line)) and a match line ML (a third word line) to be a second port.

Since a row decoder 16 , a control circuit 31 and column selectors 21 and 22 are the same as those in the first embodiment shown in FIG. 1 , description will be omitted.

A control circuit 33 controls operations of a driver circuit 26 and a coincidence detecting circuit 27 under timing control of an add comparison control input signal ME.

The driver circuit 26 inputs a data input signal DI 2 <m−1: 0 > corresponding to m search line pairs SL and bar SL, that is, expectation data for one row to the CAM cell array 13 at time of data comparison and retrieval under the control of the control circuit 33 .

The coincidence detecting circuit 27 outputs match output data MO and hit output data BO based on signal values of a plurality of match lines ML in the CAM cell array 13 at time of data comparison and retrieval under the control of the control circuit 33 .

(Memory Cell Structure)

FIG. 12 is a circuit diagram showing a memory cell array structure of the 1-port memory cell array 11 and the CAM cell array 13 . For convenience of explanation, FIG. 12 shows only a memory cell having a 2×2 structure but does not imply an actual size of the memory cell array (a matrix-like memory cell arrangement).

Since the 1-port memory cell array 11 is the same as the structure of the first embodiment shown in FIG. 2 , description will be omitted.

On the other hand, the CAM cell array 13 is constituted by memory cells mc 00 , mc 01 , mc 10 and mc 11 .

Each of the memory cells mc 00 to mc 11 is constituted by inverters 41 and 42 which are cross connected to each other, an NMOS transistor Q 11 having one of electrodes connected to an input of the inverter 41 , an NMOS transistor Q 12 having one of electrodes connected to an input of the inverter 42 , and NMOS transistors Q 23 to Q 26 .

The NMOS transistors Q 23 and Q 24 having gate electrodes connected to the inputs of the inverters 42 and 41 respectively have mutual electrodes connected to each other at a node N 23 , and one of electrodes of each of the NMOS transistors Q 25 and Q 26 to which a gate electrode of the node N 23 is connected in common is grounded.

In the memory cells mc 00 and mc 01 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to a word line WL 1 < 0 > for the first port in common, and the other electrodes of the NMOS transistors Q 25 and Q 26 are connected to a match line ML< 0 > in common.

In the memory cells mc 10 and mc 11 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to a word line WL 1 < 1 > for the first port in common, and the other electrodes of the NMOS transistors Q 25 and Q 26 are electrically connected to a match line ML< 1 > in common.

Both of the memory cells mc 00 and mc 10 on the same column are provided between a bit line pair BL 21 < 0 > and bar BL 21 < 0 > for the first port. In these memory cells mc 00 and mc 10 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 21 < 0 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 21 < 0 > for the first port.

Furthermore, both of the memory cells mc 00 and mc 10 are provided between a search line pair SL< 0 > and bar SL< 0 >. In these memory cells mc 00 and mc 10 , the other electrode of the NMOS transistor Q 23 is electrically connected to the search line SL< 0 > and the other electrode of the NMOS transistor Q 24 is electrically connected to the inverted search line bar SL< 0 >.

Both of the memory cells mc 01 and mc 11 on the same column are provided between a bit line pair BL 21 < 1 > and bar BL 21 < 1 > for the first port. In these memory cells mc 01 and mc 11 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 21 < 1 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 21 < 1 > for the first port.

Furthermore, both of the memory cells mc 01 and mc 11 are provided between a search line pair SL< 1 > and bar SL< 1 >. In these memory cells mc 01 and mc 11 , the other electrode of the NMOS transistor Q 23 is electrically connected to the search line SL< 1 > and the other electrode of the NMOS transistor Q 24 is electrically connected to an inverted search line bar SL< 2 >.

The match lines ML< 0 > and ML< 1 > are amplified by drivers 56 and 55 (which are usually provided in the coincidence detecting circuit 27 (not shown)), respectively.

(Operation)

With reference to FIGS. 11 and 12 , an operation for comparing and retrieving data through the second port will be described below. Since reading and writing operations using the first port are the same as those in the first embodiment except that a 2-port memory cell array 12 is replaced with the CAM cell array 13 , description will be omitted.

The operation for comparing and retrieving data through the second port will be described below. For convenience of the description, it is assumed that memory cell array structures of the 1-port memory cell array 11 and the CAM cell array 13 are set to be n (rows)×m (columns)=2×2.

In the case in which the operation for comparing and retrieving the data through the second port is to be executed, all of the match lines ML< 0 > and ML< 1 > are precharged to “H” and a comparison control input signal ME is then enabled. At the same time, expectation data transferred from the outside are given as a data input signal DI 2 < 1 : 0 > to the driver circuit 26 . Accordingly, all of the match lines ML< 0 > and ML< 1 > are brought into a selection state.

The driver circuit 26 drives each of the search line pairs SL< 0 > and bar SL< 0 > and SL< 1 > and bar SL< 1 > to “H” or “L” based on the data input signal DI 2 < 1 : 0 >. As a result, the expectation data and stored data can be compared for all the memory cells mc 00 to mc 11 of the CAM cell array 13 .

The details of the comparing operation will be described below. There will be supposed the case in which expectation data of “1” are given to the search line pair SL< 0 > and bar SL< 0 > (“H” or “L” is given to the search line pair SL< 0 > and bar SL< 0 > ).

In this case, if the stored contents of the memory cell mc 00 is “1” (an output of the inverter 41 is “L” (an output of the inverter 42 is “H”)), the NMOS transistor Q 23 is turned OFF and the NMOS transistor Q 24 is turned ON so that “L” is propagated to the node N 23 . As a result, the NMOS transistors Q 25 and Q 26 are maintained to be OFF and the match line ML< 0 > is maintained to be “H” (coincidence).

On the other hand, if the stored contents of the memory cell mc 00 is “0” (the output of the inverter 41 is “H” (the output of the inverter 42 is “L”)), the NMOS transistor Q 23 is turned ON and the NMOS transistor Q 24 is turned OFF so that “H” is propagated to the node N 23 . As a result, the NMOS transistors Q 25 and Q 26 are maintained to be ON and the match line ML< 0 > is led into “L” (non-coincidence).

The same comparing operation is carried out between the memory cell mc 10 and the match line ML< 1 > which are provided between the search line pair SL< 0 > and bar SL< 0 > and between the memory cells mc 01 and mc 11 and the match lines ML< 0 > and ML< 1 > which are provided between the search line pair SL< 1 > and bar SL< 1 >.

In the case in which results of the comparison in both of the memory cells mc 00 and mc 01 are coincident with each other, accordingly, the match line ML< 0 > is set to be “H”. On the other hand, in the case in which the results of the comparison are not coincident with each other, the match line ML< 0 > is set to be “L”.

Similarly, in the case in which results of the comparison in both of the memory cells mc 10 and mc 11 are coincident with each other, the match line ML< 1 > is set to be “H”. On the other hand, in the case in which the results of the comparison are not coincident with each other, the match line ML< 1 > is set to be “L”.

Signals obtained from the match lines ML< 0 > and ML< 1 > are amplified by drivers 56 and 55 respectively and signals of all the match lines ML are output as 2-bit match output data M 0 by the coincidence detecting circuit 27 which is not shown in FIG. 12 , and furthermore, an AND of the signals of all the match lines ML is output as 1-bit output data BO. More specifically, when all the match lines ML are “H” (coincident), the hit output data BO are set to be“H”. In other cases, the hit output data BO are set to be “L”.

As described above, it is possible to read and write 8-bit data which are divided and held in high and low orders on a 4-bit unit in the 1-port memory cell array 11 and the CAM cell array 13 from the first port.

Then, it is possible to carry out the data comparing operation for lower 4-bit data which are held in the CAM cell array 13 from the second port. In this case, the upper 4-bit data are not accessed at all.

As a result, the semiconductor storage according to the fourth embodiment can divide, on a 4-bit unit, data to be accessed on an 8-bit length unit and can access data on a lower 4-bit unit thus obtained by the division (the comparing operation).

Thus, the semiconductor storage according to the fourth embodiment can control the 1-port memory cell array 11 and the CAM cell array 13 including memory cells having different circuit structures from each other by the row decoder 16 . Therefore, it is possible to obtain a semiconductor storage having a 1-chip structure in which a single port memory cell and a CAM memory cell having different structures can be accessed at the same time.

(Layout Structure)

FIG. 13 is a circuit diagram showing the details of a CAM memory cell MC. As shown in FIG. 13 , the inverter 41 is constituted by a CMOS inverter including a PMOS transistor P 1 and an NMOS transistor N 1 and the inverter 42 is constituted by a CMOS inverter including a PMOS transistor P 2 and an NMOS transistor N 2 .

Other structures are the same as those described above in relation to the memory cell mc 00 in FIG. 12 or the like except that the word line WL, the bit line pair BL and bar BL, the match line ML, and the search line pair SL and bar SL are generalized.

FIGS. 14 to 16 are explanatory views showing a layout structure of the CAM memory cell MC. FIG. 14 is an explanatory view showing a layout structure in all layers as seen on a plane. FIG. 15 is an explanatory view mainly showing a layout structure provided under a first aluminum wiring layer as seen on a plane. FIG. 16 is an explanatory view showing a layout structure provided above a second aluminum wiring layer as seen on a plane. In FIG. 14 , the designations are partially omitted in consideration of easiness of understanding.

As shown in FIG. 15 , the PMOS transistors P 1 and P 2 are formed in an N well region NW, the NMOS transistor N 1 and the NMOS transistors Q 23 to Q 26 are formed in a P well region PW 1 , and the NMOS transistor N 2 and the NMOS transistors Q 11 and Q 12 are formed in a P well region PW 0 . The P well region PW 0 and the P well region PW 1 are formed opposite to each other with the N well region NW interposed therebetween.

In the N well region NW, the PMOS transistor P 1 is constituted by a polysilicon wiring PL 1 provided across P + diffusion regions FL 110 and FL 101 and the PMOS transistor P 2 is constituted by a polysilicon wiring PL 2 provided across P + diffusion regions FL 100 and FL 111 .

In the P well region PW 1 , the NMOS transistors N 1 and Q 23 are constituted by the polysilicon wiring PL 1 provided across N + diffusion regions FL 201 and FL 211 and FL 202 and FL 213 , and the NMOS transistor Q 24 is constituted by the polysilicon wiring PL 2 provided across N + diffusion regions FL 230 and FL 213 . The polysilicon wiring PL 1 is formed from the N well region NW to the P well region PW 1 and is thus shared as gates of the NMOS transistors N 1 and Q 23 and the PMOS transistor P 1 .

Furthermore, the NMOS transistor Q 25 is constituted by an upper side section PL 5 a of the polysilicon wiring PL 5 provided across the N + diffusion regions FL 203 and FL 214 , and the NMOS transistor Q 26 is constituted by a lower side section PL 5 b of the polysilicon wiring PL 5 provided across the N + diffusion regions FL 214 and FL 231 . The polysilicon wiring PL 5 is formed to be cross U-shaped in which two sides (PL 5 a and PL 5 b ) are provided in parallel with each other, and is shared by the NMOS transistors Q 25 and Q 26 .

In the P well region PW 0 , the NMOS transistor N 2 is constituted by the polysilicon wiring PL 2 provided across N + diffusion regions FL 200 and FL 210 and the NMOS transistors Q 11 and Q 12 are constituted by a polysilicon wiring PL 3 provided across N + diffusion regions FL 210 and FL 220 , and FL 212 and FL 221 , respectively. The polysilicon wiring PL 2 is formed from the P well region PW 1 to the P well region PW 0 through the N well region NW and is thus shared as gates of the NMOS transistor Q 24 , the PMOS transistor P 2 and the NMOS transistor N 2 , and the polysilicon wiring PL 3 is shared as gates of the NMOS transistors Q 11 and Q 12 .

Moreover, the polysilicon wiring PL 1 , the polysilicon wiring PL 3 and the upper side section PL 5 a of the polysilicon wiring PL 5 are formed on the same straight line, and the polysilicon wiring PL 2 and the lower side section 5 b of the polysilicon wiring PL 5 are formed on the same straight line.

As a result, the PMOS transistor P 1 and the NMOS transistors N 1 , Q 11 , Q 12 , Q 23 and Q 25 can be formed along the same straight line and the PMOS transistor P 2 and the NMOS transistors N 2 , Q 24 and Q 26 can be formed along the same straight line. Consequently, a cell height CHC of the CAM memory cell MC can be set corresponding to two transistors in the same manner as the single port memory cell MS (CHC=HC 1 ).

In the P well region PW 0 , a ground wiring LG 1 provided on the N + diffusion region FL 200 is electrically connected to the diffusion region FL 200 through a diffusion contact hole CH. The polysilicon wiring PL 3 is electrically connected to a word line 1 WL 1 through a gate contact hole GC, a bit line 1 BL 21 provided on the diffusion region FL 220 is electrically connected to the diffusion region FL 220 through the diffusion contact hole CH, and the inverted bit line bar 1 BL 21 provided on the diffusion region FL 221 is electrically connected to the diffusion region FL 221 through the diffusion contact hole CH.

Furthermore, the polysilicon wiring PL 1 and the N + diffusion region FL 212 are electrically connected to each other through the shared contact SC formed from the N + diffusion region FL 212 to the polysilicon wiring PL 1 .

In the N well region NW, a power wiring LV 1 provided on the diffusion region FL 100 is electrically connected to the diffusion region FL 100 through the diffusion contact hole CH, and the power wiring LV 1 provided on the diffusion region FL 101 is electrically connected to the diffusion region FL 101 through the diffusion contact hole CH.

In the P well region PW 1 , the search line 1 SL (the first layer aluminum wiring) provided on the diffusion region FL 230 is electrically connected to the diffusion region FL 230 through the diffusion contact hole CH. The inverted search line bar 1 SL (the first layer aluminum wiring) provided on the diffusion region FL 202 is electrically connected to the diffusion region FL 202 through the diffusion contact hole CH.

The aluminum wiring AL 105 to be the first layer aluminum wiring extended from the diffusion region FL 213 to the polysilicon wiring PL 5 is electrically connected to the diffusion region FL 213 through the diffusion contact hole CH 1 , and is electrically connected to the polysilicon wiring PL 5 through the gate contact hole GC.

The ground wiring LG 1 provided on the diffusion region FL 201 is electrically connected to the diffusion region FL 201 through the diffusion contact hole CH, the ground wiring LG 1 provided on the diffusion region FL 203 is electrically connected to the diffusion region FL 203 through the diffusion contact hole CH, and the ground wiring LG 1 provided on the diffusion region FL 231 is electrically connected to the diffusion region FL 231 through the diffusion contact hole CH.

The match line 1 ML (the first layer aluminum wiring) provided on the diffusion region FL 214 is electrically connected to the diffusion region FL 214 through the diffusion contact hole CH.

The aluminum wiring AL 102 to be the first layer aluminum wiring which is extended from the diffusion region FL 210 of the P well region PW 0 to the diffusion region FL 111 of the N well region NW is electrically connected to the diffusion region FL 210 through the diffusion contact hole CH. The aluminum wiring AL 102 is electrically connected to the diffusion region FL 111 and the polysilicon wiring PL 1 through the shared contact SC formed from the diffusion region FL 111 to the polysilicon wiring PL 1 .

The aluminum wiring AL 103 to be the first layer aluminum wiring which is extended from the diffusion region FL 211 of the P well region PW 1 to the diffusion region FL 110 of the N well region NW is electrically connected to the diffusion region FL 211 through the diffusion contact hole CH. The aluminum wiring AL 103 is electrically connected to the diffusion region FL 110 and the polysilicon wiring PL 2 through the shared contact SC formed from the diffusion region FL 110 to the polysilicon wiring PL 2 .

As shown in FIG. 16 , a ground wiring LG 2 is electrically connected to the lower ground wiring LG 1 (not shown) through a via hole VH 1 . A power wiring LV 2 is electrically connected to the lower power wiring LV 1 (not shown) through the via hole VH 1 .

A word line 2 WL 1 is electrically connected to the word line 1 WL 1 (not shown) through the via hole VH 1 , and a word line 3 WL 1 is electrically connected to the word line 2 WL 1 through a via hole VH 2 . These word lines 1 WL 1 to 3 WL 1 constitute the word line WL 1 for the first port in FIGS. 11 to 13 .

A match line 2 ML is electrically connected to the match line 1 ML (not shown) through the via hole VH 1 , and a word line 3 ML is electrically connected to the match line 2 ML through the via hole VH 2 . These match lines 1 ML to 3 ML constitute the match line ML for the second port in FIGS. 11 to 13 .

The word lines 3 WL 1 and 3 ML are formed across the P well regions PW 0 and PW 1 and the N well region NW.

The bit line 2 BL 21 is electrically connected to the lower bit line 1 BL 21 (not shown) through the via hole VH 1 and the inverted bit line bar 2 BL 21 is electrically connected to the lower inverted bit line bar 1 BL 21 (not shown) through the via hole VH 1 .

The search line 2 SL (the second layer aluminum wiring) is electrically connected to the lower search line 1 SL (not shown) through the via hole VH 1 . The inverted search line bar 2 SL (the second layer aluminum wiring) is electrically connected to the lower inverted search line bar 1 SL through the via hole VH 1 .

The bit line pair BL 21 and bar BL 21 for the first port in FIGS. 11 to 13 are constituted by the bit line 2 BL 21 , the bit line 1 BL 21 and the inverted bit lines bar 2 BL 21 and bar 1 BL 21 , and the search line pair SL and bar SL in FIGS. 11 to 13 are constituted by the search lines 2 SL and 1 SL and the inverted search lines bar 2 SL and bar 1 SL.

A pass wiring 2 DL (a second layer aluminum wiring) is formed between the inverted search line bar 2 SL and the match line 2 ML, and is not electrically connected to any wiring. The pass wiring 2 DL does not need to be always formed.

The bit lines 2 BL 21 and bar 2 BL 21 , the search lines 2 SL and bar 2 SL, the ground wiring LG 2 , the power wiring LV 2 and the pass wiring 2 DL are formed in parallel with each other in a longitudinal direction of the drawing over the P well regions PW 1 and PW 0 and the N well region NW, respectively.

Thus, a cell height of the CAM memory cell MC is set to be equal to that of the single port memory cell MS (HC 1 =CHC). Consequently, the single port memory cell MS and the CAM memory cell MC can be arranged adjacently without a useless region such as a spacer. In this case, the word line 1 WL 1 , the gate contact hole GC and the via hole VH 1 provided on a boundary line can be shared in the same manner as in the first embodiment shown in FIG. 7 .

(Others)

It is possible to obtain the same effects as those in the third embodiment in the case in which a row decoder corresponding to the row decoder 18 is provided between the 1-port memory cell array 11 and the CAM cell array 13 as in the third embodiment in place of the row decoder 16 .

<Fifth Embodiment>

(Whole Structure)

FIG. 17 is a block diagram showing a structure of a semiconductor storage according to a fifth embodiment of the present invention,

As shown in FIG. 17 , a 2-port memory cell array 14 (a first memory cell array) and a 2-port memory cell array 12 are provided together over one chip to constitute a memory macro. More specifically, the 2-port memory cell array 14 is provided with a word line WL 1 for a first port (a first word line (a common word line)) and a word line WL 3 for a second port, and the 2-port memory cell array 12 is provided with a word line WL 1 for a first port (a second word line (a common word line)) and a word line WL 2 for a second port (a third word line).

Since row decoders 16 and 17 , control circuits 31 and 32 , and column selectors 21 to 23 are the same as those in the first embodiment shown in FIG. 1 , description will be omitted. For convenience of description, names of a bit line pair for the first port to be controlled by the column selector 21 are changed into BL 11 and bar BL 11 .

Upon receipt of an address input bus signal AD 3 , a control circuit 34 supplies a row address to a row decoder 18 and a column address to a column selector 24 under timing control of a read control input signal RE 3 and a write control input signal WE 3 .

A row decoder 19 drives the word lines WL 3 for the second port in the 2-port memory cell array 14 based on the row address received by the control circuit 34 .

The column selector 24 selects a portion equivalent to 4 bits from a plurality of bit line pairs BL 12 and bar BL 12 for the second port based on the column address sent from the control circuit 34 and inputs/outputs a data input/output bus signal DIO 3 < 3 : 0 > to/from the 2-port memory cell array 14 through the bit line pair BL 12 and bar BL 12 for the second port thus selected.

The column selector 24 has a sense amplifier (SA) function and a write driving

(WD) function.

(Memory Cell Structure)

FIG. 18 is a circuit diagram showing a memory cell array structure of the 2-port memory cell array 14 and the 2-port memory cell array 12 . For convenience of explanation, FIG. 18 shows only a memory cell having a 2×2 structure but does not imply an actual size of the memory cell array (a matrix-like memory cell arrangement). Since the 2-port memory cell array 12 is the same as that in the first embodiment shown in FIG. 2 , description will be omitted.

The 2-port memory cell array 14 is constituted by memory cells mw 00 , mw 01 , mw 10 and mw 11 .

Each of the memory cells mw 00 to mw 11 is constituted by inverters 41 and 42 which are cross connected to each other, NMOS transistors Q 11 and Q 31 having electrodes connected to an input of the inverter 41 , and NMOS transistors Q 12 and Q 32 having electrodes connected to an input of the inverter 42 .

Word lines WL 3 < 0 > and WL 3 < 1 > for the second port are driven by drivers 58 and 57 (which are usually provided in the row decoder 19 (not shown)), respectively.

In the memory cells mw 00 and mw 01 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 < 0 > for the first port in common, and gate electrodes of the NMOS transistors Q 31 and Q 32 are electrically connected to the word line WL 3 < 0 > for the second port in common.

In the memory cells mw 10 and mw 11 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 < 1 > for the first port in common, and gate electrodes of the NMOS transistors Q 31 and Q 32 are electrically connected to the word line WL 3 < 1 > for the second port in common.

Both of the memory cells mw 00 and mw 10 on the same column are provided between a bit line pair BL 11 < 0 > and bar BL 11 < 0 > for the first port. In these memory cells mw 00 and mw 10 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 11 < 0 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 11 < 0 > for the first port.

Furthermore, both of the memory cells mw 00 and mw 10 are provided between a bit line pair BL 12 < 0 > and bar BL 12 < 0 > for the second port. In these memory cells mw 00 and mw 10 , the other electrode of the NMOS transistor Q 31 is electrically connected to the bit line BL 12 < 0 > for the second port and the other electrode of the NMOS transistor Q 32 is electrically connected to the inverted bit line bar BL 12 < 0 > for the second port.

Both of the memory cells mw 01 and mw 11 on the same column are provided between a bit line pair BL 11 < 1 > and bar BL 11 < 1 > for the first port. In these memory cells mw 01 and mw 11 , the other electrode of the NMOS transistor Q 11 is electrically connected to the bit line BL 11 < 1 > for the first port and the other electrode of the NMOS transistor Q 12 is electrically connected to the inverted bit line bar BL 11 < 1 > for the first port.

Furthermore, both of the memory cells mw 01 and mw 11 are provided between a bit line pair BL 12 < 1 > and bar BL 12 < 1 > for the second port. In these memory cells mw 01 and mw 11 , the other electrode of the NMOS transistor Q 31 is electrically connected to the bit line BL 12 < 1 > for the second port and the other electrode of the NMOS transistor Q 32 is electrically connected to the inverted bit line bar BL 12 < 1 > for the second port.

(Operation)

Since an operation for carrying out read/write from/to the 2-port memory cell array 14 through the second port is the same as that in the 2-port memory cell array 12 , description will be omitted. Since other operations are the same as those of the first embodiment, description will be omitted.

In the semiconductor storage according to the fifth embodiment, therefore, the 2-port memory cell array 14 and the 2-port memory cell array 12 can be controlled by the row decoder 16 . Consequently, it is possible to obtain a semiconductor storage having a 1-chip structure in which the 2-port memory cells in different memory cell arrays can be accessed at the same time.

(Comparison with Second Embodiment)

In the case in which a comparison with a structure in implementation of a semiconductor storage equivalent to that of the second embodiment is carried out by using two memory macros having a 2-port memory cell structure, the numbers of the row decoders and the control circuits can be reduced by one (both of the numbers is three in FIG. 17 and both of the numbers is four in application of the second embodiment). Thus, the degree of integration can be more enhanced than that of the second embodiment.

(Layout Structure)

The 2-port memory cell array 14 is caused to have the same layout structure as that of the 2-port memory cell array 12 according to the first embodiment shown in FIGS. 5 and 6 . By setting cell heights of both of the memory cells to be equal to each other, consequently, it is possible to provide the 2-port memory cell array 12 and the 2-port memory cell array 14 adjacently without a useless region such as a spacer.

(Others)

In the case in which a row decoder corresponding to the row decoder 18 is provided between the 2-port memory cell array 12 and the 2-port memory cell array 14 as in the third embodiment in place of the row decoder 16 , it is possible to obtain the same effects as those of the third embodiment.

<Sixth Embodiment>

(Whole Structure)

FIG. 19 is a block diagram showing a structure of a semiconductor storage according to a sixth embodiment of the present invention,

As shown in FIG. 19 , a 1-port memory cell array 11 and a 2-port memory cell array 36 (a second memory cell array) are provided together over one chip to constitute a memory macro. More specifically, the 1-port memory cell array 11 is provided with a word line WL 1 for a first port (a first word line (a common word line)) and the 2-port memory cell array 36 is provided with a word line WL 1 for a first port (a second word line (a common word line)) and a word line WL 2 for a second port (a third word line).

Since row decoders 16 and 17 , a control circuit 31 , and column selectors 21 and 22 are the same as those in the first embodiment shown in FIG. 1 , description will be omitted.

Upon receipt of an address input bus signal AD 2 , a control circuit 35 supplies a row address to the row decoder 17 and a column address to a column selector 37 under timing control of a read control input signal RE 2 .

The column selector 37 selects a portion equivalent to 4 bits from a plurality of read bit line pairs RBL for the second port based on the column address sent from the control circuit 35 and inputs/outputs a data output signal DO 2 < 3 : 0 > to/from the 2-port memory cell array 36 through the read bit line pair RBL thus selected.

The column selector 37 has a sense amplifier function and a drive function.

(Memory Cell Structure)

FIG. 20 is a circuit diagram showing a structure of a 2-port memory cell 38 on one unit in the 2-port memory cell array 36 .

As shown in FIG. 20 , an inverter 41 is constituted by a CMOS inverter including a PMOS transistor P 1 and an NMOS transistor N 1 and an inverter 42 is constituted by a CMOS inverter including a PMOS transistor P 2 and an NMOS transistor N 2 .

A node Na to be an input of the inverter 41 (an output of the inverter 42 ) is connected to a bit line pair BL 21 for the first port through an NMOS transistor N 3 , and a node Nb to be an input of the inverter 42 is connected to an inverted bit line bar BL 21 for the first port through an NMOS transistor N 4 . Gate electrodes of the NMOS transistors N 3 and N 4 are connected to a word line WL 1 for the first port.

An NMOS transistor N 5 has a gate electrode connected to the node Na, a source electrode grounded, and a drain electrode connected to a read bit line RBL through an NMOS transistor N 6 . The NMOS transistor N 6 has a gate electrode connected to a word line WL 2 for a second port.

(Operation)

With reference to FIGS. 19 and 20 , description will be given to a reading operation using the second port. For convenience of explanation, a memory cell array structure of the 2-port memory cell array 36 is set to be n (rows)×m (columns). Since reading and writing operations using the first port are the same as those of the first embodiment, description will be omitted.

After all the read bit lines RBL are precharged to “H”, a read control input signal RE 2 is enabled. When the read control input signal RE 2 is brought into the enable state, the row decoder 17 for receiving a row address corresponding to an address input bus signal AD 2 through the control circuit 35 selectively sets, into an active state, a word line WL 2 <i> (i=0 to n−1) for the second port in word lines WL 2 <−1: 0 > for the second port having n bits.

Consequently, the NMOS transistors Q 21 and Q 22 of each of selected memory cells of the 2-port memory cell array 36 connected to the word line WL 2 <i> for the second port which is set into the active state are turned ON, and contents stored in each of the selected memory cells are propagated to the corresponding read bit line RBL.

For example, when a 2-port memory cell 38 is a “0” storage (the node Na is assumed to be “H”), both the NMOS transistor N 5 and the NMOS transistor N 6 selected by the word line WL 2 for the second port are turned ON. Consequently, the read bit line RBL is set to be “L”.

On the other hand, when the 2-port memory cell 38 is a “1” storage (the node Na is assumed to be “L”), the NMOS transistor N 5 is turned OFF. Consequently, the read bit line RBL is maintained to be “H” even if the NMOS transistor N 6 selected by the word line WL 2 for the second port is turned ON.

Thus, the contents stored in the 2-port memory cell 38 are read through all the read bit lines RBL.

The column selector 37 for receiving the column address corresponding to the address input bus signal AD 2 through the control circuit 35 selects four read bit lines from m read bit line pairs RBL, and senses and amplifies them by means of an internal sense amplifier. Consequently, the data output signal DO 2 < 3 : 0 > is output as 4-bit read data.

As described above, it is possible to read and write 8-bit data which are divided and held in high and low orders on a 4-bit unit in the 1-port memory cell array 11 and the 2-port memory cell array 36 from the first port.

Then, it is possible to read lower 4-bit data which are held in the 2-port memory cell array 36 from the second port. In this case, the upper 4-bit data are not accessed at all.

As a result, the semiconductor storage according to the sixth embodiment can divide, on a 4-bit unit, data to be accessed on an 8-bit length unit and can access data on a lower 4-bit unit thus obtained by the division. The lower 4-bit is accessible from the first and second ports so that it is possible to carry out reading to meet a demand for a data dividing access.

In the semiconductor storage according to the sixth embodiment, thus, the 1-port memory cell array 11 and the 2-port memory cell array 36 including memory cells having different circuit structures from each other can be controlled by the row decoder 16 . Therefore, it is possible to obtain a semiconductor storage having a 1-chip structure in which a single port memory cell and a 2-port memory cell having different structures can be accessed at the same time.

(Comparison with Second Embodiment)

In the case in which a comparison with a structure in implementation of a semiconductor storage equivalent to that of the sixth embodiment is carried out by using the memory macro having the 1-port memory cell structure and the memory macro having the 2-port memory cell structure as in the second embodiment, the numbers of the row decoders and the control circuits can be reduced by one (both of the numbers is two in FIG. 17 and both of the numbers is three in the structure according to the second embodiment) in the same manner as in the first embodiment. Thus, the degree of integration can be more enhanced than that of the second embodiment.

(Layout Structure)

FIGS. 21 to 23 are explanatory views showing a layout structure of a 2-port memory cell 38 in the 2-port memory cell array 36 . FIG. 21 is an explanatory view showing a layout structure in all layers as seen on a plane. FIG. 22 is an explanatory view mainly showing a layout structure provided under a first aluminum wiring layer as seen on a plane. FIG. 23 is an explanatory view showing a layout structure provided above a second aluminum wiring layer as seen on a plane. In FIG. 21 , the designations are partially omitted in consideration of easiness of understanding.

As shown in FIG. 22 , the PMOS transistors P 1 and P 2 are formed in an N well region NW, the NMOS transistor N 1 and the NMOS transistors N 5 and N 6 are formed in a P well region PW 1 , and the NMOS transistor N 2 and the NMOS transistors N 3 and N 4 are formed in a P well region PW 0 . The P well region PW 0 and the P well region PW 1 are formed opposite to each other with the N well region NW interposed therebetween.

In the N well region NW, the PMOS transistor P 1 is constituted by a polysilicon wiring PL 1 provided across P + diffusion regions FL 110 and FL 101 and the PMOS transistor P 2 is constituted by a polysilicon wiring PL 2 provided across P + diffusion regions FL 100 and FL 111 .

In the P well region PW 1 , the NMOS transistors N 1 and N 5 are constituted by the polysilicon wiring PL 1 provided across N + diffusion regions FL 201 and FL 211 and FL 213 and FL 201 and the NMOS transistor N 6 is constituted by a polysilicon wiring PL 4 provided across N + diffusion regions FL 230 and FL 213 . The polysilicon wiring PL 1 is formed from the N well region NW to the P well region PW 1 and is thus shared as gates of the NMOS transistors N 1 and N 5 and the PMOS transistor P 1 .

In the P well region PW 0 , the NMOS transistor N 2 is constituted by the polysilicon wiring PL 2 provided across N + diffusion regions FL 200 and FL 210 and the NMOS transistors N 3 and N 4 are constituted by a polysilicon wiring PL 3 provided across N + diffusion regions FL 210 and FL 220 , and FL 212 and FL 221 , respectively. The polysilicon wiring PL 2 is formed from the N well region NW to the P well region PW 0 and is thus shared as gates of the NMOS transistor N 2 and the PMOS transistor P 2 , and the polysilicon wiring PL 3 is shared as gates of the NMOS transistors N 3 and N 4 .

Moreover, the polysilicon wiring PL 1 and the polysilicon wiring PL 3 are formed on the same straight line, and the polysilicon wiring PL 2 and the polysilicon wiring PL 4 are formed on the same straight line.

As a result, the PMOS transistor P 1 and the NMOS transistors N 1 and N 3 to N 5 can be formed along the same straight line and the PMOS transistor P 2 and the NMOS transistors N 2 and N 6 can be formed along the same straight line. Consequently, a cell height CH 3 of the 2-port memory cell 38 can be set corresponding to two transistors in the same manner as the single port memory cell MS (CH 3 =HC 1 ).

The P + diffusion regions FL 1 . . . are obtained by implanting and diffusing a P-type impurity and the N + diffusion regions FL 2 . . . are obtained by implanting and diffusing an N-type impurity.

In the P well region PW 0 , a ground wiring LG 1 provided on the N + diffusion region FL 200 is electrically connected to the diffusion region FL 200 through a diffusion contact hole CH. The polysilicon wiring PL 3 is electrically connected to a word line 1 WL 1 through a gate contact hole GC. A bit line 1 BL 21 provided on the diffusion region FL 220 is electrically connected to the diffusion region FL 220 through the diffusion contact hole CH, and the inverted bit line bar 1 BL 21 provided on the diffusion region FL 221 is electrically connected to the diffusion region FL 221 through the diffusion contact hole CH.

Furthermore, the polysilicon wiring PL 1 and the N + diffusion region FL 212 are electrically connected to each other through the shared contact SC formed from the N + diffusion region FL 212 to the polysilicon wiring PL 1 .

In the N well region NW, a power wiring LV 1 provided on the diffusion region FL 100 is electrically connected to the diffusion region FL 100 through the diffusion contact hole CH, and the power wiring LV 1 provided on the diffusion region FL 101 is electrically connected to the diffusion region FL 101 through the diffusion contact hole CH.

In the P well region PW 1 , the read bit line 1 RBL (the first layer aluminum wiring) provided on the diffusion region FL 230 is electrically connected to the diffusion region FL 230 through the diffusion contact hole CH.

The word line 1 WL 2 provided on the polysilicon wiring PL 4 is electrically connected to the polysilicon wiring PL 4 through the gate contact hole GC, the ground wiring LG 1 provided on the diffusion region FL 201 is electrically connected to the diffusion region FL 201 through the diffusion contact hole CH, and the ground wiring LG 1 provided on the diffusion region FL 202 is electrically connected to the diffusion region FL 202 through the diffusion contact hole CH.

The aluminum wiring AL 102 to be the first layer aluminum wiring which is extended from the diffusion region FL 210 of the P well region PW 0 to the diffusion region FL 111 of the N well region NW is electrically connected to the diffusion region FL 210 through the diffusion contact hole CH. The aluminum wiring AL 102 is electrically connected to the diffusion region FL 111 and the polysilicon wiring PL 1 through the shared contact SC formed from the diffusion region FL 111 to the polysilicon wiring PL 1 .

The aluminum wiring AL 103 to be the first layer aluminum wiring which is extended from the diffusion region FL 211 of the P well region PW 1 to the diffusion region FL 110 of the N well region NW is electrically connected to the diffusion region FL 211 through the diffusion contact hole CH. The aluminum wiring AL 103 is electrically connected to the diffusion region FL 110 and the polysilicon wiring PL 2 through the shared contact SC formed from the diffusion region FL 110 to the polysilicon wiring PL 2 .

As shown in FIG. 23 , a ground wiring LG 2 is electrically connected to the lower ground wiring LG 1 (not shown) through a via hole VH 1 . A power wiring LV 2 is electrically connected to the lower power wiring LV 1 (not shown) through the via hole VH 1 .

A word line 2 WL 1 is electrically connected to the word line 1 WL 1 (not shown) through the via hole VH 1 , and a word line 3 WL 1 is electrically connected to the word line 2 WL 1 through a via hole VH 2 . These word lines 1 WL 1 to 3 WL 1 constitute the word line WL 1 for the first port in FIGS. 19 and 20 .

A word line 2 WL 2 is electrically connected to the word line 1 WL 2 (not shown) through the via hole VH 1 , and a word line 3 WL 2 is electrically connected to the word line 2 WL 2 through the via hole VH 2 . These word lines 1 WL 2 to 3 WL 2 constitute the word line WL 2 for the second port in FIGS. 19 and 20 .

The word lines 3 WL 1 and 3 WL 2 are formed across the P well regions PW 0 and PW 1 and the N well region NW.

The bit line 2 BL 21 (the second layer aluminum wiring) is electrically connected to the lower bit line 1 BL 21 (not shown) through the via hole VH 1 and the inverted bit line bar 2 BL 21 (the second layer aluminum wiring) is electrically connected to the lower inverted bit line bar 1 BL 21 (not shown) through the via hole VH 1 .

The bit line 2 RBL (the second layer aluminum wiring) is electrically connected to the lower bit line RBL (not shown) through the via hole VH 1 .

The bit line pair BL 21 and bar BL 21 for the first port in FIGS. 19 and 20 are constituted by the bit line 2 BL 21 , the bit line 1 BL 21 and the inverted bit lines bar 2 BL 21 and bar 1 BL 21 , and the read bit line RBL in FIGS. 19 and 20 is constituted by the bit line 2 RBL and the bit line 1 RBL.

The bit lines 2 BL 21 and bar 2 BL 21 , the read bit line 2 RBL, the ground wiring LG 2 and the power wiring LV 2 are formed in parallel with each other in a longitudinal direction of the drawing over the P well regions PW 1 and PW 0 and the N well region NW, respectively.

Thus, a cell height of the 2-port memory cell 38 is set to be equal to that of the single port memory cell MS (HC 1 =CH 3 ). Consequently, the single port memory cell MS and the 2-port memory cell 38 can be arranged adjacently without a useless region such as a spacer. In this case, the word line 1 WL 1 , the gate contact hole GC and the via hole VH 1 provided on a boundary line can be shared in the same manner as in the first embodiment shown in FIG. 7 .

(Others)

It is possible to obtain the same effects as those in the third embodiment in the case in which a row decoder corresponding to the row decoder 18 is provided between the 1-port memory cell array 11 and the 2-port memory cell array 36 as in the third embodiment in place of the row decoder 16 .

<Seventh Embodiment>

(Whole Structure)

FIG. 24 is a block diagram showing a structure of a semiconductor storage according to a seventh embodiment of the present invention,

As shown in FIG. 24 , 1-port memory cell arrays 15 a (a first memory cell array) and 15 b (a third memory cell array) for timing regulation are formed on both ends to interpose a 2-port memory cell array 15 m (a second memory cell array) for a real operation formed in a central part therebetween. These memory cell arrays 15 a , 15 b and 15 m are formed integrally. A plurality of dummy cells in at least one column are arranged over a plurality of rows in the 1-port memory cell arrays 15 a and 15 b for timing regulation.

Upon receipt of an address input bus signal AD 1 , a control circuit 30 A supplies a row address to a row decoder 20 A and a column address to a column selector 25 A under timing control of a read control input signal RE 1 and a write control input signal WE 1 .

Upon receipt of an address input bus signal AD 2 , a control circuit 30 B supplies a row address to a row decoder 20 B and a column address to a column selector 25 B under timing control of a read control input signal RE 2 and a write control input signal WE 2 .

The row decoder 20 A (the first row decoder) drives a plurality of word lines WL 1 for the first port (first and second word lines (common word lines)) based on the row address, and the row decoder 20 B drives a plurality of word lines WL 2 for the second port (a third word line) based on the row address.

The column selector 25 A selects a predetermined bit from a plurality of bit line pairs BL 1 and bar BL 1 for the first port based on the column address sent from the control circuit 30 A, and inputs/outputs a data input/output bus signal DIOA to/from the 1-port memory cell array 11 through the bit line pairs BL 1 and bar BL 1 for the first port thus selected.

The column selector 25 B selects a predetermined bit from a plurality of bit line pairs BL 2 and bar BL 2 for the second port based on the column address sent from the control circuit 30 B, and inputs/outputs a data input/output bus signal DIOB to/from the 2-port memory cell array 12 through the bit line pairs BL 2 and bar BL 2 for the second port thus selected.

These column selectors 25 A and 25 B have an SA function, a WD function and a dummy sense function which will be described below.

(Memory Cell Structure)

FIG. 25 is a circuit diagram showing a structure of a dummy cell 6 ( 6 a , 6 b ) formed in 1-port memory cell arrays 15 a and 15 b for timing regulation.

As shown in FIG. 25 , an inverter 43 is constituted by a CMOS inverter including a PMOS transistor P 11 and an NMOS transistor N 11 and an inverter 44 is constituted by a CMOS inverter including a PMOS transistor P 12 and an NMOS transistor N 12 .

An input section of the inverter 44 (gates of the PMOS transistor P 12 and the NMOS transistor N 12 ) is fixed at a supply voltage.

The dummy cell 6 is constituted by an NMOS transistor N 13 having one of electrodes connected to an input (a node Na) of the inverter 43 and an NMOS transistor N 14 having one of electrodes connected to an input (a node Nb) of the inverter 44 .

Gate electrodes of the NMOS transistors N 13 and N 14 are electrically connected to the word line WL (WL 1 and WL 2 ) in common, the other electrode of the NMOS transistor N 13 is electrically connected to a dummy DBL, and the other electrode of the NMOS transistor N 14 is electrically connected to an inverted dummy bit line bar DBL.

FIG. 26 is a circuit diagram showing a memory cell array structure of a 2-port memory cell array 15 m for a real operation and the 1-port memory cell arrays 15 a and 15 b for timing regulation. For convenience of explanation, FIG. 26 shows only a memory cell having a 2×2 structure in the 2-port memory cell array 15 m for the real operation but does not imply an actual size of the memory cell array (a matrix-like memory cell arrangement).

The row decoder 20 A selectively sets one of the word lines WL 1 for the first port to be “H”. The row decoder 20 B selectively sets one of the word lines WL 2 for the second port to be “H”.

The 2-port memory cell 5 is constituted by inverters 41 and 42 which are cross connected to each other, NMOS transistors Q 11 and Q 21 having electrodes connected to an input of the inverter 41 , and NMOS transistors Q 12 and Q 22 having electrodes connected to an input of the inverter 42 .

In the 2-port memory cell 5 on the same row, gate electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the word line WL 1 for the first port in common and gate electrodes of the NMOS transistors Q 21 and Q 22 are electrically connected to the word line WL 2 for the second port in common.

In the 2-port memory cell 5 on the same column, the other electrodes of the NMOS transistors Q 11 and Q 12 are electrically connected to the common bit lines BL 1 and bar BL 1 for the first port respectively and the other electrodes of the NMOS transistors Q 21 and Q 22 are electrically connected to the common bit lines BL 2 and bar BL 2 for the second port respectively.

In the dummy cell 6 a of the 1-port memory cell array 15 a for timing regulation, the gate electrodes (first dummy ports) of the NMOS transistors N 13 and N 14 are electrically connected to the word line WL 1 for the first port in common and the other electrodes of the NMOS transistors N 13 and N 14 are electrically connected to dummy bit lines DBL 0 and bar DBL 1 respectively.

In the dummy cell 6 b of the 1-port memory cell array 15 b for timing regulation and the gate electrodes (second dummy ports) of the NMOS transistors N 13 and N 14 are electrically connected to the word line WL 2 for the second port in common and the other electrodes of the NMOS transistors N 13 and N 14 are electrically connected to dummy bit lines DBL 1 and bar DBL 1 respectively.

In the semiconductor storage according to the seventh embodiment, thus, the 1-port memory cell arrays 15 a and 15 b for timing regulation are provided on both ends of the 2-port memory cell array 15 m for the real operation.

(Timing Regulating Operation)

A timing regulating operation to be carried out by the dummy cell 6 with such a structure will be described for the dummy cell 6 a of the 1-port memory cell array 15 a for timing regulation.

In execution of a first port reading operation of the 2-port memory cell array 15 m for the real operation and the 1-port memory cell array 15 a for timing regulation, when the dummy bit line pair DBL and bar DBL are precharged to be “H” and the word line WL 1 for the first port thus selected is then set to be “H”, an electric potential of a dummy bit line DBL 0 of the dummy cell 6 a connected to the word line WL 1 for the first port thus selected is extracted into “L” and an electric potential of an inverted dummy bit line bar DBL 0 is held to be “H”.

In this case, the fact that an electric potential of the dummy bit line DBL extracted to be the “L” level is detected by a dummy sense function in the column selector 25 A (a voltage level detecting circuit or a dummy sense circuit) is set to be a trigger, and a sense amplifier activating signal for activating the SA function for the 2-port memory cell array 15 m for the real operation is thus brought into an active state. As a result, it is possible to carry out a stable reading operation for the 2-port memory cell 5 of the 2-port memory cell array 15 m for the real operation.

Accordingly, a transistor size of the dummy cell 6 a is adapted to the 2-port memory cell 5 and is thus optimized. Thus, a desirable read timing can be obtained.

By carrying out the timing regulating operation in the same manner for the dummy cell 6 b , similarly, it is possible to produce the same effects in the reading operation of the second port.

Thus, the dummy cell 6 a is used for generating a reading timing of the first port and the dummy cell 6 b is used for generating a reading timing of the second port. Consequently, it is possible to regulate the reading timing without a hindrance by using the dummy cells 6 a and 6 b having a smaller number of ports than those of the 2-port memory cell 5 .

Since an operation for reading two ports in the 2-port memory cell array 15 m for the real operation is the same as an operation for existing two ports and the 2-port memory cell, description will be omitted.

(Layout Structure)

A 1-port memory cell having a different port structure from that of the 2-port memory cell 5 provided in the 2-port memory cell array 15 m for the real operation is used for the dummy cell 6 . In general, a cell size of the 2-port memory cell has an area which is equal to an approximately double of an approximately double area of the 1-port memory cell. This is also apparent from a comparison of the layout structure of the memory cell structure (see FIGS. 3 , 4 and 7 ) of the 1-port memory cell array 11 according to the first embodiment with the layout structure of the memory cell structure (see FIGS. 5 to 7 ) of the 2-port memory cell array 12 , for example.

As a result, the semiconductor storage according to the seventh embodiment comprises the 1-port memory cell arrays 15 a and 15 b for timing regulation including the dummy cell 6 having a 1-port structure adjacently to the 2-port memory cell area 15 m for the real operation. Consequently, it is possible to produce an advantage that formation areas of the 1-port memory cell arrays 15 a and 15 b for timing regulation can be reduced, and furthermore, the degree of integration of the whole apparatus can be enhanced.

More specifically, in the semiconductor storage according to the seventh embodiment, the number of the ports that the dummy cell can access is set to be smaller than that of the ports that the 2-port memory cell 5 to be the memory cell for the real operation can access. Consequently, it is possible to enhance the degree of integration of the apparatus with a reduction in the formation area.

While the invention has been shown and described in detail, the foregoing description is in all aspects illustrative and not restrictive. It is therefore understood that numerous modifications and variations can be devised without departing from the scope of the invention.