WO2010113713A1 - 半導体装置 - Google Patents

半導体装置 Download PDF

Info

Publication number
WO2010113713A1
WO2010113713A1 PCT/JP2010/055029 JP2010055029W WO2010113713A1 WO 2010113713 A1 WO2010113713 A1 WO 2010113713A1 JP 2010055029 W JP2010055029 W JP 2010055029W WO 2010113713 A1 WO2010113713 A1 WO 2010113713A1
Authority
WO
WIPO (PCT)
Prior art keywords
mlut
wiring
address
circuit
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2010/055029
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
哲夫 弘中
一哉 谷川
博昭 戸口
直樹 平川
隆 石黒
正幸 佐藤
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
HIROSHIMA CITY
Taiyo Yuden Co Ltd
Original Assignee
HIROSHIMA CITY
Taiyo Yuden Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by HIROSHIMA CITY, Taiyo Yuden Co Ltd filed Critical HIROSHIMA CITY
Priority to CN201080013413.3A priority Critical patent/CN102369668B/zh
Priority to US13/255,846 priority patent/US8283945B2/en
Publication of WO2010113713A1 publication Critical patent/WO2010113713A1/ja
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03K—PULSE TECHNIQUE
    • H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
    • H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
    • H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
    • H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
    • H03K19/17736—Structural details of routing resources
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03K—PULSE TECHNIQUE
    • H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
    • H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
    • H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
    • H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
    • H03K19/17724—Structural details of logic blocks
    • H03K19/17728—Reconfigurable logic blocks, e.g. lookup tables
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03K—PULSE TECHNIQUE
    • H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
    • H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
    • H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
    • H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
    • H03K19/1778—Structural details for adapting physical parameters
    • H03K19/17792—Structural details for adapting physical parameters for operating speed

Definitions

  • the present invention relates to a configuration of a programmable semiconductor device including a memory and including a logic function cell having an address line and a data line pair.
  • a semiconductor device equipped with an LSI Large Scale Integration
  • LSI Large Scale Integration
  • a typical example of a programmable semiconductor device has a variable function circuit cell group, a variable signal wiring group, a switch element group, and a memory. Based on the information of the memory, the switch by the transistor is turned on / off (ON / OFF). Then, there is an FPGA (Field Programmable Gate Array) that switches between the variable function circuit cell group and the variable signal wiring group to realize a circuit having a desired function in a programmable manner.
  • FPGA Field Programmable Gate Array
  • each circuit and each signal wiring are changed using a switch using a transistor.
  • the occupied area of the circuit and the wiring is fixed, there are few degrees of freedom, and depending on a desired functional circuit, In some cases, the configuration becomes useless. Therefore, the area efficiency in the layout is generally poor, an LSI having a relatively large chip area is required, and the cost performance is low. Also, the manufacturing process is long and complicated including the number of wiring layers, and the manufacturing cost is high. Further, in the conventional FPGA, when the wiring connecting each circuit extends to a distant position, transmission delay due to parasitic capacitance and parasitic resistance becomes a problem, which may cause malfunction.
  • Patent Document 1 Patent Document 1
  • Patent Document 2 Patent Document 2
  • the MPLD is configured as follows. That is, a memory having the same number of address lines as the number of data lines is mounted, and data for forming a desired function is written into the memory. An element that performs an equivalent function to various circuits by arranging an MLUT having a function of reading output data corresponding to an operation from the memory as a unit cell, arranging unit cells of the plurality of MLUTs side by side, and wiring each other. What is obtained is MPLD.
  • MPLD is an abbreviation for Memory-based Programmable Logic Device
  • MLUT is an abbreviation for MPLD Look Up Table.
  • the wiring connection between the flip-flop and the MLUT is inefficient because the flip-flop is disposed outside the MPLD.
  • problems such as a long connection wiring, a long delay time, and a high speed cannot be achieved.
  • the MLUT is used as a circuit element resource and is converted to a flip-flop or a latch circuit, and then the logic that requires the flip-flop using the circuit is used.
  • the present invention solves such problems, and its object is to provide high functionality and stable characteristics, and has a high degree of design freedom, good chip area efficiency, and thus high cost performance. It is an object of the present invention to provide a programmable semiconductor device, in particular, a semiconductor device comprising an MPLD that is improved with respect to the aforementioned problems.
  • each invention is configured as follows.
  • an MLUT block including N (N is an integer of 1 or more) MLUT blocks each having a memory, an address line, and a data line pair is used as a basic unit, and a plurality of MLUT blocks that are the basic units are provided. Are arranged, and flip-flops are arranged in the MLUT block.
  • the second invention is the one in which a D-type flip-flop (Delay-type flip-flop) is used as the flip-flop in the first invention.
  • two or more flip-flops are arranged in the MLUT block.
  • At least two flip-flops are arranged in the MLUT block.
  • the memory is composed of SRAM (Static Random Access Memory).
  • the memory is a non-volatile memory.
  • short-distance wirings by pairs of address lines and data lines between the MLUTs are alternately arranged wirings.
  • short-distance wirings formed by pairs of address lines and data lines between the MLUTs are alternately arranged wirings.
  • a ninth invention is any one of the first to fourth inventions, or the seventh invention or the eighth invention, wherein a separate wiring is provided for signal connection between the MLUTs or between MLUT blocks. It is.
  • a tenth aspect of the invention is the ninth aspect of the invention, wherein the separated wiring includes a torus structure.
  • the flip-flop is provided exclusively in the MLUT block, not only a combinational circuit but also a sequential circuit can be configured. Therefore, there is an effect that a high-speed and high-performance circuit can be easily configured. In addition, since area efficiency and element efficiency are improved, there is an effect that it can be provided at a lower cost.
  • FIG. 6 is a wiring pattern diagram showing two types of short-distance wiring patterns of MLUTs that perform alternately arranged wiring in the MLUT block according to the embodiment of the present invention. It is the circuit block diagram which showed the terminal arrangement
  • FIG. 6 is a wiring pattern diagram showing two types of short-distance wiring patterns of MLUTs that perform alternately arranged wiring in the MLUT block according to the embodiment of the present invention. It is the circuit block diagram which showed the terminal arrangement
  • FIG. 3 is a circuit block diagram showing a configuration of a plurality of address lines and data lines of an MLUT according to an embodiment of the present invention, and input address lines and input / output data lines during a memory operation.
  • FIG. 3 is a circuit block diagram showing a relationship among MLUTs, flip-flops, and short-distance wirings, in which MLUT blocks according to an embodiment of the present invention are regularly arranged on a plane. It is a functional diagram showing the operation function of the full adder circuit of the conventional circuit, and a circuit diagram showing a circuit configuration example. It is a circuit block diagram which shows the structure of the 4-bit addition circuit using MLUT concerning the Example of this invention.
  • FIG. 1 is a circuit block diagram showing a first example of the configuration of an MLUT block including nine MLUTs which are main elements of the semiconductor device of the present invention.
  • M0 to M8 are MLUTs.
  • MLUT blocks are configured by MLUTM0 to M8.
  • MLUTM0 to M8 form a circuit by exchanging signals composed of a pair of address lines and data lines (hereinafter also abbreviated as “address-data pair”) between MLUTs adjacent to each other on the same plane.
  • Wiring between MLUTs adjacent to each other on the same plane is referred to as short-distance wiring.
  • wiring between MLUTs that are not adjacent to each other on the same plane is referred to as separated wiring.
  • FIG. 1 only short-distance wirings with a pair of address lines and data lines, which are features of the present invention, are shown with numbers, but other wirings are not shown in order to avoid confusion.
  • the signal wiring not shown in FIG. 1 will be described later with reference to FIGS.
  • the MLUTM4 is provided with a D-type (Delay) flip-flop.
  • a D-type flip-flop is sometimes abbreviated as D-FF in the drawings.
  • FIG. 4 is a circuit block diagram showing a specific configuration of the MLUT which is a basic unit circuit constituting the semiconductor device of the present invention. The configuration and operation will be described below.
  • the memory block 40 is configured by arranging memory cells made of SRAM (Static Random AccessnMemory) in a matrix of n ⁇ (2 to the power of n) (n is a positive integer). Further, the address lines 51 of the memory block 40 are (2 to the power of n), the number of data inputs 44 is n, and the number of data outputs 52 is n.
  • SRAM Static Random AccessnMemory
  • the input operation switching signal 49 is a memory operation signal.
  • the input address (n) 45 is fetched from the address switching circuit 41 during memory operation, the address decoder 42 designates an n-th power address, and only the n SRAM cells in a row corresponding to the address are designated in the memory block 40.
  • the data of the expected value necessary for the address is stored in the SRAM cell in the column through the n signal lines from the data input 44 during the input operation.
  • the logic operation is selected by the operation switching signal 49.
  • the input address (n) 47 during logic operation is fetched from the address switching circuit 41, and only n SRAM cells corresponding to a desired address selected via the address decoder 42 are activated.
  • N pieces of data are taken out to the data lines (n pieces) 52, and taken out as output data to the data output (n pieces) 48 in the logic operation through the output data switching circuit 43. Since this output data is the expected value of the output with respect to the input value as the desired logic circuit, it functions equivalently as a logic circuit. If the address as the input value changes, another output value corresponding to the input value is obtained. Therefore, an operation equivalent to a desired logic circuit is performed for a predetermined input value.
  • the MLUT is operated as a logic circuit, but it can also be used as a simple memory circuit.
  • the operation switching signal 49 is selected for the memory operation, the addresses are sequentially selected from the input address 45 during the memory operation, and the n SRAM cells of the memory block 40 are activated via the address decoder 42, and the memory Write n data from the data input 44 during operation. This operation is performed for each address, and the write operation is completed.
  • the memory operation is selected by the operation switching signal 49.
  • the memory switching input address 45 is fetched from the address switching circuit 41, and only n SRAM cells corresponding to the desired address 51 selected via the address decoder 42 are activated, and the data line 52 of the memory block 40 is activated.
  • N data are taken out from the output data switching circuit 43 and taken out as output data to the memory operation data output (n pieces) 46.
  • memory operation data input 44 and the memory operation data output 46 are not used at the same time, and may be used as bidirectional input / output lines.
  • either function can be assigned to the MLUT as a logic circuit or a memory circuit.
  • the memory block 40 is composed of SRAM, the memory data disappears once the power is turned off. Therefore, when the power is turned on again, the data is not rewritten to the memory block 40 as desired. The circuit does not operate.
  • FIG. 5 is a circuit block diagram showing a wiring configuration for further clarifying the relationship between FIG. 4 showing the configuration in the MLUT and FIG. 1 showing the connection relationship of the short-distance wiring of the MLUT.
  • both the memory block 40 and the memory operation input address 45 correspond to the memory block 40 and the memory operation input address 45 of FIG.
  • the memory operation input / output data 4446 shown in FIG. 5 corresponds to the memory operation data input 44 and the memory operation data output 46 shown in FIG.
  • Each book is divided separately and corresponds to a reconfiguration of seven pairs of input address lines and data output lines.
  • the address switching circuit 41, the address decoder 42, the output data switching circuit 43, and the operation switching signal 49 in FIG. 4 are not shown in order to avoid confusion between expression and understanding.
  • the pairs P0 to P5 correspond to the address-data pairs P0 to P5 in FIG. Yes.
  • the address-data pairs P0 to P5 are represented by a single line. However, as shown in FIG. 5, the address-data pair is actually composed of two lines.
  • the remaining address-data pair P6 in FIG. 5 is used for a separate wiring not shown in FIG.
  • MLUTM0 to M8 are each composed of the above-mentioned MLUT, and have various circuit functions depending on the memory data written in the memory block in the MLUT. These MLUTM0 to M8 are connected to each other using address-data pairs P0 to P5, which are short-distance wiring, and the address corresponds to the input signal and the data line corresponds to the output signal. More complex circuits are constructed.
  • the D-type flip-flop 11 is specially arranged in the MLUTM4.
  • the signal wiring connection between the MLUTM4 and the D-type flip-flop 11 uses an address-data pair P6 that is not used for the short-distance wiring.
  • the address-data pair P6 is used for connection to the D-type flip-flop 11, and therefore the separated wiring is not used only for MLUTM4.
  • MLUTM0 to M3 and MLUTM5 to M8 are provided with separate wiring terminals instead of being accompanied by D-type flip-flops, and perform separate wiring using the remaining address-data pair P6 in FIG. It exchanges signals with circuits located far away.
  • MLUTM0 and MLUTM3 are arranged so that the address-data pairs P0 to P5 come out in an upside down manner.
  • address-data pairs P1 and P2 of the first type MLUT represented by M0 are arranged in the left direction
  • address-data pairs P3 and P4 are arranged in the right direction.
  • Each is used as a short distance wiring with MLUT located in the horizontal direction.
  • the address-data pair P0 is located in the lower left direction
  • the address-data pair P5 is located in the lower right direction, and is used as a short-distance wiring between the MLUTs located on the left and right sides of the lower stage, respectively.
  • the configuration of the address-data pair arrangement of the first type MLUT is as described above.
  • address-data pairs P1 and P2 of the second type MLUT represented by M3 are arranged in the right direction, and address-data pairs P3 and P4 are arranged in the left direction.
  • Each is used as a short distance wiring with MLUT located in the horizontal direction.
  • the address-data pair P0 is located in the upper right direction, and the address-data pair P5 is located in the upper left direction, and is used as a short-distance wiring between the MLUTs located on the upper right side and the left side, respectively.
  • the arrangement of the address-data pair arrangement of the second type MLUT is as described above.
  • the first-type MLUT represented by M0 and the second-type MLUT represented by M3 have a symmetrical configuration with respect to the arrangement of the address-data pairs used as short-distance wiring in the vertical and horizontal directions. Therefore, as shown in FIG. 1, the MLUT has an efficient configuration for connecting an address-data pair as a short-distance wiring between the MLUTs close to each other. In this way, two types of MLUTs in which the positional relationship of the address-data pair shown in FIG. 2 is symmetrical in the left-right and top-bottom relationship are alternately arranged in one direction, and based on this, short-range placement and routing by the address-data pair What has been performed is referred to as alternate wiring.
  • FIG. 1 the above-described alternate arrangement and wiring composed of two types of MLUTs are given as examples of efficient arrangement and wiring.
  • address-data pairs are arranged in the vertical direction of M0, M1, and M2 of the MLUT.
  • various other MLUT layouts in which address-data pairs are arranged in the vertical direction and arrangements of address-data pairs are also conceivable.
  • the address-data pair P6 used for the remote wiring is used.
  • the memory operation input address 45 and the memory operation input / output data 4446 actually exist.
  • These wirings not shown in FIG. 1 are mainly arranged in the vertical direction in FIG.
  • the vertical short distance wiring is not necessarily required. In general, it cannot be said that efficient arrangement and wiring are used. Therefore, considering the presence of wiring other than the short distance wiring arranged in the vertical direction, it can be said that the alternately arranged wiring shown in FIG. 1 is an excellent configuration as the short distance wiring.
  • the alternate arrangement wiring regarding the arrangement of the MLUT, two types of symmetrical left and right MLUTs are arranged on the layout pattern from the layout pattern in which each element including the transistor is arranged, and the address-data pair. There are two ways in which two types of symmetrical left and right MLUT wiring are used only for the wiring layer mainly composed of the short distance wiring layer.
  • FIG. 6 shows a state in which MLUT blocks composed of the nine MULTs of FIG. 1 are further repeatedly arranged.
  • a region indicated by a broken line 601 represents the MLUT block in FIG. MLUTs 611 and 612 and flip-flops 621 are also regularly arranged outside the MLUT indicated by a broken line 601.
  • FIG. 6 for the sake of simplicity, only short-distance wiring based on address-data pairs is shown, and separated wiring, MLUT control lines, and the like are not shown.
  • a circuit having a desired function is configured by combining the plurality of MLUT blocks.
  • a semiconductor device is formed by providing a plurality of MLUT blocks.
  • FIG. 1 shows a state in which MLUTM0 to M8 are arranged in an orderly grid
  • FIG. 6 shows a state in which MLUTM0 to M8 are arranged in a cluster almost located at the apex of the rhombus, but in FIG.
  • FIG. 8 shows an example in which a 4-bit adder is realized using an MLUT block group.
  • FIG. 7A shows a diagram representing a truth value as a 1-bit full adder.
  • A is the operand
  • B is the operand
  • C is the carry from the previous bit
  • S is the sum operation result
  • Ca is the carry to the next bit.
  • the calculation result S and the carry Ca to the next bit are determined. It expresses what will happen.
  • FIG. 7B is an example of a standard circuit that realizes the relationship between the above truth values. Since FIG. 7B is a generally well-known circuit configuration, detailed description thereof is omitted.
  • one MLUT having the structure shown in FIG. 4 is assigned the inputs A, B, and C of FIG. 7A to the input address 45 at the time of memory operation, and outputs S, Ca to the data input 44 at the time of memory operation. And the respective combinations of inputs A, B, and C are input to the input address 45 during memory operation, and the output S, corresponding to the combination is input to the activated 2-bit SRAM cell specified by the address at that time.
  • Each value of Ca is written to the activated 2-bit SRAM cell through the data input 44 during memory operation.
  • the MLUT has a function as a 1-bit full adder.
  • each of the outputs S and Ca as the full adder from the logical operation data output 48 is input.
  • the value is output.
  • the logical input address 47 and the logical operation data output 48 are included in six address-data pairs P0 to P5 as shown in FIGS. 1 and 5, and are adjacent to each other as short-distance wirings.
  • the MLUT can exchange input / output signals.
  • FIG. 8 is a circuit block diagram in which a 4-bit addition circuit is configured in the MLUT block group.
  • MLA1, MLA2, MLA3, and MLA4 are MLUTs each having the function of a 1-bit full adder circuit by the above-described method.
  • MLS1, MLS2, MLS3, and MLS4 are MLUTs each having a function in which an input value and an output value are the same, and have a function corresponding to an ON state of the switch.
  • An input signal A1 and B1 corresponding to the operation value of the first bit and the operation value are input from the input address of the input address-data pair to the MLA1 having a full adder function.
  • the operation result of the first bit is output as S1 from the output data of the input address-data pair.
  • An MLA2 having a full adder circuit function receives a second bit operand value and input signals A2 and B2 corresponding to the operand value from the input address of the input address-data pair.
  • the operation result of the second bit is output as S2 from the output data of the input address-data pair.
  • An MLA3 having a full adder circuit function receives the third bit operand value and input signals A3 and B3 corresponding to the operand value from the input address of the input address-data pair.
  • the operation result of the third bit is output as S3 from the output data of the input address-data pair.
  • An MLA4 having a full adder circuit function receives the fourth bit operand value and input signals A4 and B4 corresponding to the operand value from the input address of the input address-data pair.
  • the operation result of the fourth bit is output as S4 from the output data of the input address-data pair.
  • the MLS3 having a wiring switch function inputs Ca, which is a carry to the next bit, from the full adder MLA3, and outputs C through the input address-data pair as C, which is a carry from the previous bit of the full adder MLA4.
  • the MLS 4 having a wiring switch function receives Ca, which is a carry to the next bit, from the full adder MLA 4 and outputs it as a carry as a 4-bit adder.
  • the input signal is a 4-bit (A1, A2, A3, A4) operation value and the 4-bit (B1, B2, B3, B4) operation value is the input signal, which is the operation result (S1 S2, S3, S4) and a carry output signal Ca as a carry signal.
  • circuits using flip-flops In the above, an example of application to the 1-bit full adder circuit in FIG. 7 and the 4-bit adder circuit in FIG. 8 is shown, but all are so-called combinational circuits that are uniquely determined by the input signal at that time. On the other hand, even if the input signal is the same, a circuit whose output result differs depending on a prior state is called a sequential circuit.
  • the sequential circuit generally requires a latch circuit or a flip-flop circuit that stores the previous state.
  • FIG. 14 is a circuit block diagram showing in more detail the connection relationship between MULTM4 and the D-type flip-flop in FIG.
  • the address-data pair P6 in FIG. 5 is used between the MLUT 1401 and the D-type flip-flop 1402.
  • the address-data pair P6 is an address-data pair that is left for the exchange of signals with the D-type flip-flop.
  • the address-data pair P6 of the MLUT 1401 is connected to the D terminal 1411 and the Q terminal 1412 of the D-type flip-flop 1402.
  • a signal serving as a reference for synchronization is selected and connected to the CL signal terminal 1413 of the D-type flip-flop 1402.
  • FIGS. 15A and 15B are layout block diagrams showing the arrangement relationship between each MLUT and the D-type flip-flop.
  • a plurality of MLUTs 1561 to 1563 and the like are arranged, and a plurality of D-type flip-flop groups 1521 are arranged. This shows a state where a large number of D-type flip-flops are collectively arranged outside the MLUT group.
  • FIG. 15B shows a state in which the D-type flip-flops 1512 and 1513 are widely and uniformly arranged in the MLUT group (1552 to 1553, etc.) as shown in FIGS.
  • FIG. 15A when the D-type flip-flop is used by each MLUT, the positional relationship is far away, so that there are many cases where wiring is wasted or congested and the layout is not appropriate.
  • FIG. 15B since the D-type flip-flop exists relatively close to each MLUT, it can be used efficiently. It will be understood that the MLUT block configuration according to FIG. 1 is still excellent.
  • FIG. 9 is a circuit diagram showing a first example of a specific circuit configuration of the D-type flip-flop shown in FIGS.
  • the circuit shown in FIG. 9 is a 1-bit master D flip-flop. The configuration and operation are described below.
  • the D-type flip-flop is composed of inverter circuits 904 and 902 and clocked gate inverter circuits 901 and 903.
  • the clocked gate inverter circuit is an inverter circuit (an inverting circuit) that determines whether to output an output signal depending on whether the clock signal is 1 or 0.
  • a clock signal (CL) for controlling the operation timing is input to the input terminal of the inverter circuit 904, and an inverted signal of the clock signal is output from the output terminal.
  • a data signal (D) 905 to be controlled is input to the input terminal of the clocked gate inverter circuit 901, and an inverted signal of the data signal is output to the output terminal.
  • CL clock signal
  • An input terminal of the inverter circuit 902 is connected to an output terminal of the clocked gate inverter circuit 901, and an output terminal of the inverter circuit 902 is connected to an output terminal (Q) 907 as a D-type flip-flop terminal.
  • the output terminal of the inverter circuit 902 is connected to the input terminal of the clocked gate inverter circuit 903.
  • the output terminal of the clocked gate inverter circuit 903 is connected to the input terminal of the inverter circuit 902.
  • the output of the clocked gate inverter circuit 903 is connected so as to be controlled by the output signal of the inverter circuit 904, it is controlled by the inverted signal of the clock signal (CL).
  • the signal is binary and the positive signal and the negative signal are expressed by 1 and 0, respectively
  • the clock signal (CL) input from the terminal 906 is 1, the input signal terminal 905 of the D flip-flop
  • the data signal (D) is inverted by the clocked gate inverter circuit 901 and further inverted by the inverter circuit 902, and the same signal as the data signal (D) is immediately output to the output signal terminal 907 of the D-type flip-flop.
  • the same signal as the data signal (D) of the output signal terminal 907 is input to the input terminal of the clocked gate inverter circuit 903, but the inverter circuit is an inverted signal of the clock signal (CL). Since the output signal is controlled by the output signal 904, the output signal is not output to the output terminal of the clocked gate inverter circuit 903 (neither 1 nor 0 is in a high impedance state).
  • the inverter circuit 902 and the clocked gate inverter circuit 903 constitute a latch circuit, and store and hold the data signal (D) in the previous state while the clock signal (CL) is zero. While the clock signal (CL) is 0, the data signal (D) in the previous state is held and output regardless of how the data at the input signal terminal 905 of the D-type flip-flop changes.
  • the D-type flip-flop outputs a new output signal in response to a new data signal (D) after the clock signal (CL) becomes 1. Until then, the data signal (D) in the previous state is stored and output.
  • FIG. 11 is a circuit diagram showing a second example of a specific circuit configuration of the D-type flip-flop.
  • the circuit shown in FIG. 11 is a 2-bit master-slave D-type flip-flop.
  • a 1-bit D-type flip-flop of a master unit composed of inverter circuits 904 and 902 and clocked gate inverter circuits 901 and 903, inverter circuits 1104 and 1102, and clocked gate inverter circuit 1101,
  • a 1-bit D-type flip-flop of the slave section composed of 1103.
  • the above-described master unit has a function as a 1-bit D-type flip-flop.
  • the configuration of the inverter circuits 1104 and 1102 and the clocked gate inverter circuits 1101 and 1103 in the slave portion of FIG. 11 is completely the same as that of the inverter circuits 904 and 902 and the clocked gate inverter circuits 901 and 903 in the master portion of FIG.
  • the configuration is the same, and the last digit of each number corresponds in turn. Therefore, the slave section including the inverter circuits 1104 and 1102 and the clocked gate inverter circuits 1101 and 1103 in FIG. 11 also has a function as a 1-bit D-type flip-flop.
  • the clock signal CL supplied to the clocked gate inverter circuits 1101 and 1103 and the inverted signal thereof have an opposite phase relationship between the clock signal CL supplied to the clocked gate inverter circuits 901 and 903 and the inverted signal thereof. It has become.
  • the output terminal M of the inverter 902 which is an output unit of the master unit is connected to the input terminal of the clocked gate inverter circuit 1101 which is an input terminal of the slave unit. Note that the output terminal M of the master unit input to the input terminal of the slave unit is output as the output signal Q of the slave unit to the output signal of the inverter circuit 1102 through the 1-bit D-type flip-flop of the slave unit.
  • the 2-bit master-slave type Operates as a D-type flip-flop.
  • FIG. 12 is a circuit diagram showing a third example of the specific circuit configuration of the D-type flip-flop.
  • the circuit shown in FIG. 12 is a 2-bit master-slave D-type flip-flop.
  • the D-type flip-flop of the master unit includes inverter circuits 1222 and 1223 and transmission gate circuits 1224 and 1225.
  • the transmission gate circuit has an N-type MOSFET and a P-type MOSFET connected in parallel, and each gate electrode functions as a switch that controls on / off by applying signals of opposite phases. have.
  • the MOSFET is a field effect transistor and is an abbreviation for Metal-Oxide-Semiconductor Field-Effect Transistor.
  • inverter circuits 1222 and 1223 constitute a latch circuit via a transmission gate circuit 1225. Further, it has a data transfer function (inversion of D to M in FIG. 12) and a latch function of data (inversion of M) via transmission gate circuits 1224 and 1225.
  • the D-type flip-flop of the slave unit includes inverter circuits 1232 and 1233 and transmission gate circuits 1234 and 1235.
  • the inverter circuits 1232 and 1233 constitute a latch circuit via the transmission gate circuit 1235. Further, it has a data transfer function (from M inversion to Q in FIG. 12) and a data (M inversion) latch function through transmission gate circuits 1234 and 1235.
  • inverter circuits 1204 and 1205 respectively generate an inverted signal of CL and a double inverted signal (that is, a normal rotation signal) and supply them to the transmission gate circuits 1224, 1225, 1234, and 1235.
  • FIG. 10 is a circuit diagram of a D-type flip-flop having a configuration different from that of FIG.
  • the circuit shown in FIG. 10 is a 1-bit master-type D-type flip-flop.
  • a latch circuit in which input / output lines of NAND circuits 1002 and 1003 are separated from each other, and NAND circuits 1001 and 1004 and an inverter circuit 1005 are combined, and an input data signal D and a control clock are used as D-type flip-flops.
  • a signal CL, an output signal Q, and an inverted output signal of Q are provided. Since the circuit configuration of FIG. 10 is well known, detailed description thereof is omitted.
  • FIG. 13 is a circuit diagram showing a configuration in which two 1-bit master-type D-type flip-flops of FIG. 10 are stacked to form a 2-bit master-slave-type flip-flop.
  • a circuit comprising NAND circuits 1001, 1002, 1003, 1004 and an inverter circuit 1005 is exactly the same as the circuit of FIG.
  • the configurations of NAND circuits 1301, 1302, 1303, and 1304 have basically the same functions corresponding to the NAND circuits 1001, 1002, 1003, and 1004, respectively.
  • FIG. 13 shows a 2-bit master-slave type flip-flop.
  • FIGS. 9 to 13 show various configuration examples of a 1-bit master type D-type flip-flop and a 2-bit master-slave type D-type flip-flop. Therefore, it can be understood that there are various functions and configurations even if they are called D-type flip-flops. An optimum D-type flip-flop configuration may be selected according to the application.
  • flip-flops such as an RS flip-flop, a JK flip-flop, and a T flip-flop as well as a D flip-flop. Since the function and configuration of these flip-flops are well known, detailed description thereof will be omitted. However, various flip-flops to be mounted on the MLUT block can be selected as necessary.
  • Wiring between MLUTs adjacent to each other on the same plane is called short distance wiring.
  • Wiring between MLUTs or MLUT blocks that are not adjacent to each other on the same plane is called separation wiring.
  • P0 to P5 of the address-data pair in FIG. 5 are used.
  • the address-data pair P6 of FIG. 5 is used.
  • the address-data pair P6 is used for the D-type flip-flop, but the address-data pairs P6 of MLUTM0 to M3 and MLUTM5 to M8 are all used for the separated wiring.
  • all the wirings shown in FIG. 1 are only short-distance wirings based on the address-data pairs P0 to P5, and the separated wirings based on the address-data pair P6 are not shown.
  • the short distance wiring by the address-data pair P0 to P5 is not shown, and only the separated wiring by the address-data pair P6 is shown in the wiring diagram of FIG. In FIG. 16, it is composed of nine MLUTs, MLUTM0 to M8. Only MLUTM4 is accompanied by a D-type flip-flop. However, the D-type flip-flop is not shown. As described above, since the D-type flip-flop is attached to the MLUTM4, the address-data pair P6 is used for the flip-flop and the separated wiring cannot be used.
  • Each address-data pair P6 of MLUTM0 to M3 and MLUTM5 to M8 shows a state in which the separated wiring extends to the outside of the MLUT block. Further, as described above, only MLUTM4 is different from the other MLUTM0 to M3 and MLUTM5 to M8 in that the D-type flip-flop is arranged and the separated wiring is not used. It is distinguished from the expression in which other M0 to M3 and M5 to M8 are circled.
  • FIG. 17 is a wiring diagram showing two types of patterns for connecting separated wirings.
  • FIG. 17 shows two kinds of separated wiring patterns.
  • FIG. 17A the separated wiring of the address-data pair P6 in MLUTM0, M5 is routed downward, and the separated wiring of the address-data pair P6 in MLUTM3, M8 is wired upward.
  • FIG. 17A is referred to as a separated wiring pattern 1.
  • FIG. 17B the separated wiring of the address-data pair P6 in MLUTM0, M5 is routed upward, and the separated wiring of the address-data pair P6 in MLUTM3, M8 is wired downward.
  • FIG. 17B is referred to as a separated wiring pattern 2.
  • the separated wiring pattern 2 is turned up, and the separated wiring pattern 1 is arranged.
  • the separated wiring of the address-data pair P6 in the MLUTM3 and M8 in FIG. 17B of the separated wiring pattern 2 is wired downward, and FIG. ) Of the address-data pair P6 in the MLUTM3 and M8 are routed upward, so that the MLUTM3 and M8 of the two MLUT blocks each having the separated wiring pattern 2 and the separated wiring pattern 1 are in a positional relationship. Applicable wiring is connected.
  • the spaced wiring pattern 1 and the spaced wiring pattern 2 have the same type of spaced wiring patterns with respect to MLUTM1, M2, M6, and M7. Even in the case where the two MLUT blocks of the separated wiring pattern 1 and the separated wiring pattern 2 are arranged on the left and right, in the case of FIGS. Therefore, even when the same type of separated wiring patterns are combined, it is possible to connect wires that are suitable for the positional relationship.
  • MLUT blocks are arranged in a tile shape by combining the separated wiring pattern 1 and the separated wiring pattern 2 shown in FIGS. 17A and 17B, and a part of the separated wiring via each MLUT of each MLUT block.
  • FIG. Since the purpose is to show the arrangement of separated wirings in each MLUT between the upper, lower, left and right MLUT blocks, MLUT blocks and separated wirings in other positions that may interfere with understanding may be displayed in an overlapping manner. Other MULT separation wirings that may be complicated are intentionally not shown.
  • the MLUT block 1801 is composed of the above-described separated wiring pattern 1
  • the MLUT block 1802 is composed of the separated wiring pattern 2. It can be seen that the wiring connection suitable for the positional relationship is performed in the vertical wiring of the separated wirings in the MLUTTM0 and M5 in the MLUT block 1801 and the separated wiring in the MLUTM0 and M5 in the MLUT block 1802.
  • the MLUT block 1803 is composed of the above-described separated wiring pattern 2.
  • the MLUT block 1801 includes the separated wiring pattern 1. Therefore, the wiring connection suitable for the positional relationship is performed in the vertical direction wiring between the downwardly spaced wiring in MLUTM3 and M8 in the MLUT block 1803 and the upwardly spaced wiring in MLUTM3 and M8 in the MLUT block 1801. You can see that
  • FIG. 19 is a layout layout diagram showing an efficient combination of layout patterns when MLUT blocks including separated wirings are regularly arranged on a plane.
  • FIG. 19 shows MLUT blocks arranged in the same shape as FIG. However, in the case where layout patterns are automatically repeated and arranged efficiently, the combination of MLUT blocks in FIG. 18 includes a plurality of types of separated wiring patterns, which may be cumbersome. It can take time or make mistakes.
  • FIG. 19 shows an efficient MLUT block combination by repetition with the same pattern.
  • the MLUT blocks 1801, 1802, and 1803 in FIG. 19 correspond directly to the MLUT blocks 1801, 1802, and 1803 in FIG.
  • the MLUT blocks 1801 and 1802 indicated by the broken line 1910 are included, but the MLUT block 1803 is not included, and a 4 ⁇ 4 block unit is considered.
  • the MLUT block 1802 is the separated wiring pattern 2 and the MLUT block 1803 is also the separated wiring pattern 2
  • the relationship between the 4 ⁇ 4 MLUT blocks above and below the broken line next to the broken line 1901 is the same. Since the pattern is repeated and the horizontal separated wiring pattern is the same in each MLUT block in the broken line 1901, the plane is defined with the 4 ⁇ 4 MLUT blocks indicated by the broken line in FIG. 19 as a basic unit. If the above arrangement is regularly performed, a structure in which the entire MLUT blocks are automatically arranged in a state in which the separated wirings are connected to each other in the positional relationship is completed.
  • FIG. 22 is a wiring diagram in which each cell is simply connected by one signal line.
  • cells 0 and 5 located at the end points receive only one signal line.
  • cells 1 to 4 receive two signal lines. Accordingly, the symmetry between the cells 0 and 5 located at the end points and the cells 1 to 4 located therein is lost at least in the number of received signal lines. There may be differences in characteristics between cells. In this case, the end point specificity has occurred.
  • FIG. 23 is a wiring diagram in which the 0th cell and the 5th cell at the end point are connected by an additional signal line from the state of FIG. 17 with the intention of eliminating the peculiarity of the number of signal lines at the end point as much as possible. By this method, the number of signal lines received in each cell is eliminated.
  • the length of the signal line directly connecting the 0th cell and the 5th cell at the end point is abnormally long compared to the length of the connecting line between the other cells. Therefore, there is a difference in characteristics in terms of electrical sensitivity to parasitic capacitance and noise, and symmetry between cells cannot be ensured.
  • FIG. 24 is a wiring diagram that attempts to further eliminate the specificity of the number of signal lines at the end points.
  • the leftmost cell number 0 is connected to the second cell at the position where one cell is jumped, and the second cell signal line is connected to the fourth cell at the position where one cell is jumped. Is connected.
  • the rightmost cell No. 5 has a signal line connected to the cell No. 3 at the position where one cell is jumped.
  • the third cell has a signal line connected to the first cell at a position where one cell is skipped.
  • the connection from the 4th cell to the 5th cell and the connection from the 1st cell to the 0th cell are adjacent to each other. Is taking the way.
  • Such a wiring method for wiring in a loop while jumping one by one and ensuring uniform and stable performance is generally called torus wiring.
  • FIG. 24 shows one torus wiring
  • FIG. 21 illustrates a case where a plurality of wirings are performed on an address line or the like.
  • one signal in the horizontal direction, one signal is constituted by two pairs in a loop shape, and has a torus structure while jumping one by one. A plurality of signal pairs having these structures are arranged.
  • one signal is formed in a loop shape with two pairs, and has a torus structure in which one signal jumps.
  • a plurality of signal pairs having these structures are arranged.
  • FIG. 21 shows a structure in which a plurality of signal lines are arranged in a torus structure in the horizontal direction and the vertical direction, and is generally called a torus network.
  • FIG. 18 shows the separated wiring between MLUTs and MLUT blocks of the present invention as described above, but it is also an example in which a torus structure is partially used.
  • MLUT and MLUT blocks constitute basic units, respectively, so that MUT (ML is an integer equal to or greater than 1) MLUT blocks and MLUT blocks are wired to form a separated wiring by a torus structure. Is realized.
  • a configuration in which six MLUTs are skipped or a configuration in which one MLUT block is skipped can be used. Either can be interpreted.
  • the present invention can provide optimum separated wiring with uniform and stable characteristics even for distances of different lengths.
  • FIG. 20 shows an example of a wiring diagram in which a torus wiring is applied to a separated wiring included in the present invention in which MLUT blocks are regularly arranged on a plane.
  • MLUT blocks 2001, 2002, 2003, 2004, 2005, and 2006 have the separated wiring pattern 1 or the separated wiring pattern 2 shown in FIGS. 17A and 17B as separated wirings. Therefore, the separated wiring can be performed based on the concept of the torus wiring described above.
  • one MLUT block 2003 is jumped between the MLUT block 2002 and the MLUT block 2004, and the separated wiring 2R24. Is wired. Further, one MLUT block 2005 is jumped between the MLUT block 2004 and the MLUT block 2006, and the separation wiring 2R46 is wired. The above is illustrated by wiring the right side of the MLUT blocks 2002, 2003, 2004, 2005, and 2006.
  • one MLUT block 2002 is jumped between the MLUT block 2001 and the MLUT block 2003, and the separated wiring 2L13 is wired.
  • one MLUT block 2004 is jumped between the MLUT block 2003 and the MLUT block 2005, and the separation wiring 2L35 is wired. The above shows that the left side of the MLUT blocks 2001, 2002, 2003, 2004, 2005 is wired.
  • MLUT blocks are jumped one by one for MLUT blocks 2001, 2002, 2003, 2004, 2005, 2006, and separated wirings 2R24 and 2R46 are routed on the right side, and MLUT blocks are jumped one by one.
  • separated wirings 2L13 and 2L35 are wired.
  • the torus wiring can be applied to vertical separation wiring in a state where the MLUT block is regularly arranged on a plane. It can also be seen that the torus wiring can be applied to the separation wiring in the left-right direction in a state where the MLUT block is regularly arranged on the plane. Therefore, since the torus wiring can be applied to the vertical separation wiring and the horizontal separation wiring, it can be understood that the MLUT block group of FIG. 20 forms a torus wiring network.
  • the present invention is not limited to the embodiment described above.
  • the MLUT block constituting the present invention is shown as a configuration in which nine MLUTs are combined, generally a configuration of K ⁇ L (K and L are integers of 1 or more) is possible.
  • K ⁇ L configuration a plurality of MLUTs with flip-flops can be placed in the MLUT block.
  • the basic configuration is K ⁇ L (K and L are positive integers of 1 or more), other circuit elements can be arranged in place of the MLUT in some grids.
  • FIG. 1 FIG. 2, and FIG. 5, there are seven address-data pairs, of which six (6 pairs) are used for short-distance wiring and one (one pair) is used for separated wiring.
  • the address-data pairs may be composed of a total of eight addresses P0 to P7.
  • the number beyond it may be sufficient.
  • the memory capacity in the memory block in FIG. 4 increases, but the address-data pair of the short-distance wiring in FIG. 1 can be increased by one, and FIG. Can further increase a pair of separated wirings (not shown).
  • the configuration combining MLUTs is a configuration of K ⁇ L (K and L are positive integers of 1 or more) other than 3 ⁇ 3, the number of address-data pairs is There is an appropriate number other than the above-mentioned number according to the actual situation.
  • the memory block 40 included in the MLUT has been described as the SRAM as shown in FIG. 4, but may be configured as a nonvolatile memory.
  • the memory block 40 is composed of a non-volatile memory and the MLUT of FIG. 4 has a logic circuit function, the power is cut off once, and the memory is non-volatile even when the power is turned on again. There is an effect that the set circuit function operation is started immediately. Similarly, if the memory is non-volatile, it can be used immediately after the power is turned on even when it is used as a memory circuit.
  • Non-volatile memories include EEPROM (Erasable Programmable Read Only Memory) (including FLASH type), FeRAM (Ferroelectric Random Access Memory), MRAM (Magneto-resistive Random Access Memory), and the like.
  • the MLUT has exemplified the configuration shown in FIG. 4, but is not necessarily the configuration shown in FIG.
  • the configuration of the memory block is an n ⁇ (2 to the power of n) configuration, other memory blocks may be used, and address lines and data lines may be shared in some cases.
  • FIGS. 1 and 6 the case where one flip-flop exists in FIGS. 1 and 6 has been described.
  • a plurality of flip-flops may be mounted.
  • the number of states in the previous state can be increased when used as a sequential circuit, so that a complicated sequential circuit can be realized.
  • a flip-flop is used for synchronization by a control signal, synchronization and control by a two-phase clock signal or a plurality of different control signals are possible, so that a more sophisticated circuit can be configured. .
  • the flip-flops present in FIGS. 1 and 6 are described as D-type flip-flops, and examples of other flip-flops include RS-type, JK-type, and T-type flip-flops.
  • it may be a register or a counter. If a register is installed, a pipeline type adder that performs almost simultaneous calculations in parallel without waiting for the carry from the lower digit, like the 4-bit adder circuit shown in FIG. As described above, it is possible to mount a circuit with higher speed and higher functionality.
  • FIGS. 17A and 17B two types of spaced wiring patterns 1 and 2 are described in the vertical wiring, but there are also two types of spaced wiring patterns in the horizontal wiring. . Therefore, since there are two types in the vertical direction and two types in the horizontal direction, there are at least four types of separated wiring patterns when combined. Such other separated wiring patterns may be used in combination.
  • torus wiring is applied to the separated wiring
  • torus wiring it has been described that the influence of the singularity at the end point is one method for offsetting the variation in electrical characteristics.
  • topologies other than torus topological structure.
  • These other topologies can be constructed in a pseudo manner by using multilayer wiring.
  • the pseudo structure corresponding to a spherical surface, hypercube, etc. not only eliminates the singularity at the end points described above, but also a structure that cancels out variations due to the position of the space in the manufacturing process, noise that varies depending on the position, etc. There is a possibility to have.
  • the present invention is a programmable LSI whose function can be changed at any time, and is cheaper and more stable in performance than a conventional FPGA.
  • it is less expensive than the initial MPLD, and it is easy to configure a synchronous circuit and a sequential circuit, and has high functionality and high performance. Therefore, as a programmable LSI with good cost performance, there is a possibility that it will replace the conventional FPGA as a programmable LSI at the development site of new products, a model that can be changed in many types and in small quantities, and a function change may occur. There is.

Landscapes

  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Logic Circuits (AREA)
  • Design And Manufacture Of Integrated Circuits (AREA)
PCT/JP2010/055029 2009-03-30 2010-03-24 半導体装置 Ceased WO2010113713A1 (ja)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201080013413.3A CN102369668B (zh) 2009-03-30 2010-03-24 半导体装置
US13/255,846 US8283945B2 (en) 2009-03-30 2010-03-24 Semiconductor device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2009-083786 2009-03-30
JP2009083786A JP5140029B2 (ja) 2009-03-30 2009-03-30 半導体装置

Publications (1)

Publication Number Publication Date
WO2010113713A1 true WO2010113713A1 (ja) 2010-10-07

Family

ID=42828008

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2010/055029 Ceased WO2010113713A1 (ja) 2009-03-30 2010-03-24 半導体装置

Country Status (4)

Country Link
US (1) US8283945B2 (enExample)
JP (1) JP5140029B2 (enExample)
CN (1) CN102369668B (enExample)
WO (1) WO2010113713A1 (enExample)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011162116A1 (ja) 2010-06-24 2011-12-29 太陽誘電株式会社 半導体装置
JP5890733B2 (ja) * 2012-04-09 2016-03-22 太陽誘電株式会社 再構成可能な半導体装置の配置配線方法、そのプログラム、及び配置配線装置
JP5822772B2 (ja) 2012-04-11 2015-11-24 太陽誘電株式会社 再構成可能な半導体装置
US9350357B2 (en) 2012-10-28 2016-05-24 Taiyo Yuden Co., Ltd. Reconfigurable semiconductor device
JP6250548B2 (ja) * 2012-11-20 2017-12-20 太陽誘電株式会社 再構成可能な半導体装置の論理構成方法
US9762865B2 (en) * 2013-03-15 2017-09-12 James Carey Video identification and analytical recognition system
US9425800B2 (en) 2013-04-02 2016-08-23 Taiyo Yuden Co., Ltd. Reconfigurable logic device
JP6306846B2 (ja) * 2013-09-16 2018-04-04 太陽誘電株式会社 再構成可能な論理デバイス
US9423452B2 (en) 2013-12-03 2016-08-23 Taiwan Semiconductor Manufacturing Company, Ltd. Contactless signal testing
US9923561B2 (en) * 2014-10-22 2018-03-20 Taiyo Yuden Co., Ltd. Reconfigurable device
JP6390683B2 (ja) * 2016-09-28 2018-09-19 ミツミ電機株式会社 半導体集積回路

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0220923A (ja) * 1988-07-08 1990-01-24 Fujitsu Ltd プログラマブル・ロジック・デバイス
JPH11238850A (ja) * 1998-02-19 1999-08-31 Toshiba Corp 半導体集積回路
WO2007060738A1 (ja) * 2005-11-28 2007-05-31 Taiyo Yuden Co., Ltd. 半導体装置

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3471628B2 (ja) 1998-05-12 2003-12-02 日本電信電話株式会社 書き換え可能な論理回路およびラッチ回路
US6150838A (en) * 1999-02-25 2000-11-21 Xilinx, Inc. FPGA configurable logic block with multi-purpose logic/memory circuit
US6215327B1 (en) * 1999-09-01 2001-04-10 The United States Of America As Represented By The Secretary Of The Air Force Molecular field programmable gate array
JP3517839B2 (ja) 2000-11-29 2004-04-12 日本電信電話株式会社 プログラマブルセルアレイ回路
US6331788B1 (en) * 2001-07-03 2001-12-18 The United States Of America As Represented By The Secretary Of The Air Force Simplified cellular array structure for programmable Boolean networks
US6777977B1 (en) * 2002-05-01 2004-08-17 Actel Corporation Three input field programmable gate array logic circuit configurable as a three input look up table, a D-latch or a D flip-flop
JP4226383B2 (ja) * 2003-04-23 2009-02-18 株式会社ミツトヨ 測長装置
US7102387B1 (en) * 2004-12-08 2006-09-05 The United States Of America As Represented By The Secretary Of The Air Force Periodic computation structure based on 1-input lookup tables
US7632745B2 (en) * 2007-06-30 2009-12-15 Intel Corporation Hybrid high-k gate dielectric film
US8010590B1 (en) * 2007-07-19 2011-08-30 Xilinx, Inc. Configurable arithmetic block and a method of implementing a configurable arithmetic block in a device having programmable logic
US7816946B1 (en) * 2008-01-31 2010-10-19 Actel Corporation Inverting flip-flop for use in field programmable gate arrays

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0220923A (ja) * 1988-07-08 1990-01-24 Fujitsu Ltd プログラマブル・ロジック・デバイス
JPH11238850A (ja) * 1998-02-19 1999-08-31 Toshiba Corp 半導体集積回路
WO2007060738A1 (ja) * 2005-11-28 2007-05-31 Taiyo Yuden Co., Ltd. 半導体装置

Also Published As

Publication number Publication date
CN102369668A (zh) 2012-03-07
JP5140029B2 (ja) 2013-02-06
JP2010239325A (ja) 2010-10-21
CN102369668B (zh) 2014-09-17
US20120007635A1 (en) 2012-01-12
US8283945B2 (en) 2012-10-09

Similar Documents

Publication Publication Date Title
JP5140029B2 (ja) 半導体装置
US10749528B2 (en) Stacked programmable integrated circuitry with smart memory
US5019736A (en) Programmable logic cell and array
JPH07175719A (ja) ワード長をプログラム可能なメモリ
KR950005022B1 (ko) 프로그래머블 논리회로
JP6250548B2 (ja) 再構成可能な半導体装置の論理構成方法
JP4621215B2 (ja) モジュール式ｉ／ｏバンクアーキテクチャ
JP2012511263A (ja) ３ｄマイクロアーキテクチャのシステムにおいて結合する平行プレーンメモリおよびプロセッサ
JPH10135431A (ja) 半導体装置、半導体集積回路装置、フリップフロップ回路、排他的論理和回路、マルチプレクサ及び加算器
Yang et al. Review of advanced FPGA architectures and technologies
JP2013077358A (ja) 半導体装置
US7587697B1 (en) System and method of mapping memory blocks in a configurable integrated circuit
US8434045B1 (en) System and method of providing a memory hierarchy
US20130061004A1 (en) Memory/logic conjugate system
CN101290639A (zh) 半导体集成电路以及半导体集成电路的布局方法
WO2013153852A1 (ja) 再構成可能な半導体装置の配置配線方法、そのプログラム、及び配置配線装置
US9400762B2 (en) Integrated device with memory systems accessible via basic and bypass routes
US9621159B2 (en) Reconfigurable semiconductor integrated circuit and electronic device
Jagasivamani et al. Memory-systems challenges in realizing monolithic computers
CN113626373A (zh) 一种集成芯片
CN100468314C (zh) 可编程序逻辑装置
JP2011160211A (ja) 半導体集積回路、半導体集積回路の構成変更方法
CN113626372A (zh) 一种存算一体的集成芯片
JP5699817B2 (ja) メモリセル
WO2025042587A1 (en) Assembly having a face-to-face bonded chiplet

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 201080013413.3

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10758487

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 13255846

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 10758487

Country of ref document: EP

Kind code of ref document: A1