WO2005109646A1 - 多次元のスイッチトポロジーを有する集積回路 - Google Patents
多次元のスイッチトポロジーを有する集積回路 Download PDFInfo
- Publication number
- WO2005109646A1 WO2005109646A1 PCT/JP2005/005755 JP2005005755W WO2005109646A1 WO 2005109646 A1 WO2005109646 A1 WO 2005109646A1 JP 2005005755 W JP2005005755 W JP 2005005755W WO 2005109646 A1 WO2005109646 A1 WO 2005109646A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- fpga
- dimensional
- integrated circuit
- basic blocks
- wiring
- 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
Links
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
- H03K19/1778—Structural details for adapting physical parameters
- H03K19/17796—Structural details for adapting physical parameters for physical disposition of blocks
-
- H—ELECTRICITY
- H10—SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
- H10D—INORGANIC ELECTRIC SEMICONDUCTOR DEVICES
- H10D84/00—Integrated devices formed in or on semiconductor substrates that comprise only semiconducting layers, e.g. on Si wafers or on GaAs-on-Si wafers
- H10D84/90—Masterslice integrated circuits
- H10D84/903—Masterslice integrated circuits comprising field effect technology
Definitions
- the present invention relates to a wiring architecture of a field programmable gate array (hereinafter, abbreviated as “FPGA”) whose logic function can be reconfigured, and a wiring architecture and power having a multi-dimensional switch topology.
- FPGA field programmable gate array
- the present invention relates to an integrated circuit having such wiring architecture.
- the present invention also relates to a programmable device including a plurality of basic blocks and switching elements provided in each basic block, wherein the basic blocks are connected to each other via the respective switching elements.
- Patent Document 2 Japanese Patent Laid-Open Patent Publication No. 8-509344 (Publication date: October 1, 1996)
- Patent Document 7 U.S. Pat. No. 5,883,526 (Publication: March 16, 1999)
- an FPGA is generally realized by arranging basic blocks as shown in FIG. La in a two-dimensional array as shown in FIG.
- the horizontal direction is described as the X direction
- the vertical direction is described as the y direction.
- the basic block also includes a wiring channel 106 ⁇ 107 connecting between the blocks, a switch matrix 101 for specifying a path, and a logic block 102 which is also a variable logic element and memory element power.
- a wire 105 connects between the switch matrix 101 and the logic block 102.
- a wire connecting between each basic block a wire 103 connecting between adjacent basic blocks and a wire 104 connecting to a basic block two or more ahead also exist.
- FIG. 2 is a specific example of the logic block 102, and in this example, shows an example provided with a look-up table 202 as a variable logic element and a flip-flop 203 as a storage element.
- the look-up table 202 is configured by a memory such as an SRAM.
- the output signals of the look-up table 202 and the die flip flop 203 are fed back to the input of the look-up table 202 and simultaneously output to the switch matrix 101 through the wiring 105. Further, signals from outside the basic block are also input from the switch matrix 101 to the look-up table 202 through the wiring 105.
- the input section of the look-up table 202 includes a programmable switch 201 for selecting one of the output feedback of the look-up table 202 and the decode flip-flop 203 and the basic block external force signal from the wiring 105. Is equipped. Also, in recent FPGAs, blocks where a carry calculation device is added to the logic block as shown in Figure 2, memory or multipliers, etc. There is also a block consisting of only one.
- FIG. Lb shows the topology of switch matrix 101.
- arrows indicate the presence or absence of a switch connected from the left wiring channel 106a in the X direction to each wiring channel 106b '105a' 105b.
- the two-dimensional switch matrix 101 requires switches connected to the wiring channels in three directions. Here, only the connection from the wiring channel 106a on the left side in the X direction to the other wiring channels 106b, 105a, and 105b is shown! However, the wiring channel force in the direction of misalignment is also the same. There is a switch connected to the wiring channel.
- the switch matrix 101 includes switches as shown in FIG. 3, and shorts wires connected to upper, lower, left, and right basic blocks such as the wire 103 and the wire 104 shown in FIG.
- This switch comprises a multiplexer 301. Multiplexer 301 selects lines '03' drawn from the lines connecting between the upper, lower, left, and right basic blocks connected to switch matrix 101, and some of lines 105 which are outputs from logic block 102 are input 'selected'. , And a tri-state buffer 302 for driving one of the wirings connecting between the basic blocks such as the wiring 103 and the wiring 104. By using the tri-state buffer 302, the wiring of the output destination (for example, the wiring 103) can be used as a bidirectional wiring.
- a switch configuration in which a normal buffer is used instead of the tri-state buffer 302 or a buffer is provided. The switch is controlled by a memory such as SRAM.
- the user can cause the FPGA to perform a desired operation by setting an appropriate value in the memory controlling the switch before use and the memory configuring the look-up table.
- the logic block 102 in FIG. 2 has more than a force having two logic elements, which is a combination of the look-up table 202 and the flip flop 203.
- a logic block having four logic elements is prepare.
- This is the same structure as the FPGA having the hierarchical structure found in Patent Document 6.
- the entire FPGA is considered as a super region, and the basic block as a sub region.
- the superregions are connected in a regular two-dimensional array structure, as shown in Fig. 2.
- the logic elements are connected equally as shown in Fig. 2.
- the connection from the subregions to the wiring of the superregion is the switch with each logic element. It is equally connected between the matrices. In such a structure, when the subregion is enlarged, there arises a problem that the number of switches and the delay of the internal wiring rapidly increase.
- the window frame architecture shown in Patent Document 4 and the like is also considered to be a type of hierarchical structure.
- the connection in the subregion is a two-dimensional array structure, and from the edge of the subregion Connect to the wiring channel of the super region. In this structure If the region is enlarged, the wiring of the super region will be unnecessarily long, so patent documents
- FIG. 6 An example of a basic block of a three-dimensional FPGA is shown in FIG.
- the chip horizontal direction is x
- the vertical direction is y
- the vertical direction is z.
- Wiring channels 503 and 504 are wiring channels in the x and y directions, respectively
- wiring channel 506 is a wiring channel in the z direction.
- the wiring channel 506 is connected to the upper or lower tile switch matrix 501.
- the switch matrix 501 and the logic block 502 in the basic block in FIG. 5 correspond to the switch matrix 101 and the logic block 102 in the basic block in FIG. 1, respectively.
- FIG. 6 a represents a three-dimensional view of the wiring connected to the switch matrix 501.
- FIG. 6 b shows the topology of switch matrix 501.
- the number of switches is the number of tracks in the wiring channel (the wiring channel is a bundle of wiring, and each wiring constituting the wiring channel is called a track, and the number is the number of tracks in the wiring channel).
- Proportional to The number of switches per track increases in the two-dimensional switch matrix because the force with connections in the three-directional wiring channel increases in the five directions in the three-dimensional switch matrix, but in the large-scale three-dimensional FPGA, the two-dimensional The logic density of the 3D FPGA is higher because the number of tracks required for the routing channel is much smaller than that of the FPGA.
- Non-Patent Document 1 when an FPGA of 20 k gate size is configured as a four-layer three-dimensional FPGA, the wiring delay is improved by 45 to 60%, and the logic density is improved by 20 to 40%. In addition, it is shown that the degree of improvement increases as the degree of logic integration increases, and it can be seen that the effect of three-dimensional integration is higher as the scale of FPG A increases.
- the wiring delay is increased because the wiring between basic blocks passes through many switch matrices.
- the number of switch matrices passed through increases, so the ratio of wiring delay to circuit delay also increases.
- the degree of integration can not be made very high. Furthermore, due to the technical difficulty in manufacturing the devices, the yield is reduced, and a 3D FPGA with sufficient integration has not been realized. In addition, it is difficult to dissipate heat in three-dimensional integrated circuits, and there is a limit to the number of layers that can be stacked, which makes it difficult to provide a three-dimensional FPGA with a sufficient degree of integration. For this reason, it is thought that there will never be a completely ideal three-dimensional FPGA with the same characteristics in the horizontal and vertical directions. Therefore, the current three-dimensional integrated circuit technology can provide an integrated circuit and a system having such an integrated circuit that are sufficient to cope with the expansion, multifunctionality and high speed of information of today's electronic devices. .
- the present inventors have made intensive efforts in the research and development of a wiring architecture for improving the defects in the manufacturing technology, taking advantage of the advantages such as the improvement effect of the wiring delay of the 3D FPGA and the improvement of the logic density.
- multidimensional means high dimensional of three or more dimensions.
- a semiconductor integrated circuit comprising an FPGA, and (10) a conductor integrated circuit comprising an FPGA according to (4) above
- basic blocks present at the same xy position are formed in a two-dimensional manner as one subregion
- Connections in the z direction are made, and these subregions are arranged in a two-dimensional array
- connections in the X direction are made between the basic blocks existing in the same z position of the subregions arranged in the horizontal direction.
- the basic blocks are formed as one subregion in a two-dimensional array, and connections in the X direction are made between the basic blocks arranged in the horizontal direction, and connections in the y direction are made between the basic blocks arranged in the vertical direction.
- An FPGA-powered semiconductor integrated circuit is realized by connecting the lower end of the wiring channel and the upper end of the y-direction wiring channel of the second subregion adjacent to the lower side of the first subregion, (15) above In the semiconductor integrated circuit comprising the FPGA according to 11), the right end of the wiring channel in the X direction of the first subregion and the second subregion in the X direction of the second subregion adjacent to the right of the first subregion.
- the left end of the wiring channel Semiconductor consisting of FPGA realized by connecting Integrated circuit, (16)
- a semiconductor integrated circuit comprising the FPGA according to (13), wherein the sub-regions are formed by arranging the basic blocks linearly in the vertical direction.
- an FPGA-powered semiconductor integrated circuit is realized by forming the subregions by arranging the basic blocks at least once with z-axis.
- the semiconductor integrated circuit comprising the FPGA according to (16) the lower end of the z-direction wiring channel of the first subregion and the upper end of the z-direction wiring channel of the second subregion adjacent downward.
- the present invention does not particularly limit the structure of the logic block, the connection between the logic block and the switch matrix, and the like. Further, the present invention also includes one in which a logic block is replaced by a block having a functional menu such as a memory or a multiplier.
- n-dimensional FPGA In an n-dimensional FPGA (although it is a virtual existence in the case of n force or more), basic blocks aligned in an array in each axial direction are connected by an n-dimensional grid-like wiring.
- 2n direction wiring channels are connected to the switch matrix that constitutes the basic block of the FPGA, with the positive and negative directions of each axis aligned in the first direction.
- the present invention is characterized in that an FPGA having a high dimensional switch topology is embedded in a lower order chip.
- an FPGA having a high dimensional switch topology is embedded in a lower order chip.
- the heat dissipation is similar to that of a two-dimensional FPGA, and there is no concern about the negative effects of heat storage in the three-dimensional structure.
- the logic density of the FPGA ⁇ Because it is the switch that affects the operation speed most than the metal wiring, embedding the switch topology of the 3D FPGA into a 2D chip is a pass switch between logic blocks.
- the benefits of 3D FPGAs can be realized, such as a reduction in the number of stages and a reduction in the number of switches required for the FPGA.
- an FPGA having a switch topology of four or more dimensions is considered to be advantageous.
- the present invention also includes an FPGA having a switch topology of four or more dimensions.
- the problem of an increase in wiring area when realizing a large-scale FPGA is solved, enabling realization of an FPGA with a gate size of 100 million or more, and networks such as routers, SANs and blade servers. It will be used in a wider range of applications such as 'communications field, medical use', imaging fields such as industrial imaging equipment, digital video recorders such as DVD-RW players and LCD TVs, and portable digital devices. Enable expansion of
- the number of transistors that make up the switch matrix is reduced, and logic density is greatly improved. This is due to the superior topology of the 3D FPGA switch that is not found in 3D integrated circuit technology. Therefore, by embedding the topology of the 3D FPGA switch in the 2D chip, the number of transistors that make up the switch matrix can be reduced as in the 3D FPGA. Also, by embedding the topology of the 3D FPGA switch in 2 dimensions, all vertical wiring will be expanded horizontally, and the density of metal wiring will increase.
- the density of the metal wiring has a considerable margin compared to the density of the transistor, and the metal wiring is realized by the multilayer wiring, and the density of the metal wiring will further decrease because the number of layers will continue to increase in the future. . Therefore, general Multilayer wiring can not be used effectively in two-dimensional FPGAs.
- the present invention can effectively utilize multilayer wiring.
- the present invention unlike a three-dimensional FPGA that requires three-dimensional integrated circuit technology, can be realized using conventional integrated circuit technology, and there is no manufacturing problem.
- the heat dissipation is easy when it is realized as a two-dimensional chip as one of the effective realization modes of the present invention.
- the programmable device of the present invention includes a plurality of basic blocks and switching elements provided in each of the basic blocks, and the basic blocks are connected to each other via the respective switching elements.
- the programmable device at least a part of the plurality of basic blocks are arranged in a matrix, and the basic block matrix is formed by sequentially connecting along the row direction and the column direction.
- a plurality of basic blocks are arranged, and basic blocks located at corresponding matrix positions among the plurality of basic block matrices are connected to each other, and the plurality of basic block matrices are arranged in a plane.
- the above-mentioned programmable device is a circuit capable of arbitrarily setting a logic function like an FPGA, and includes an adder, a multiplier, a memory, etc. instead of the logic block in the FPGA besides the FPGA. Circuits are included.
- the above-mentioned basic block is a basic unit constituting a programmable device, and is a switch matrix, a functional circuit, a wiring bundle connecting the above switch matrix and the above functional circuit, the above switch matrix and other basic blocks. A wiring bundle is provided to connect the switch matrix.
- the switch matrix is composed of a plurality of switches having a function of connecting or disconnecting a signal path between the lines connected to the switch matrix, and each switch is a memory element or each switch itself.
- the connection or disconnection status is set based on the storage function that the terminal has.
- the functional circuit is a circuit that performs a predetermined arithmetic function.
- the above functional circuit It includes logic blocks, adders, multipliers, memories, etc.
- the function circuit may be set as the function of the function circuit based on a memory element or a memory function of the function circuit itself.
- the functional circuit may include a basic block of a lower hierarchy hierarchically configured in a basic block to which the functional circuit belongs.
- the above-mentioned wire bundle is a set of a plurality of wires for propagating a signal.
- each basic block matrix can perform functions equivalent to basic blocks arranged in a two-dimensional array, provided in a conventional two-dimensional FPGA.
- the basic blocks located at corresponding row positions among the plurality of basic block matrices are connected to each other.
- the connection between the basic blocks can be realized via a smaller number of switch matrices compared to the conventional two-dimensional FPGA, which is the same as the conventional three-dimensional FPGA.
- the logic density can be improved.
- the plurality of basic block matrices are arranged in a plane.
- the above configuration eliminates the need to form fine vertical wires that are difficult to form, thus facilitating manufacture as compared to conventional three-dimensional FPGAs, as well as conventional methods.
- the problem of heat dissipation in a two-dimensional FPGA can also be suppressed.
- a programmable device of the present invention is a programmable device according to the above (19)
- connections between basic blocks located at corresponding matrix positions among the plurality of basic block matrices may be one-dimensional connections.
- the programmable device of the present invention is the programmable device according to the above (19), wherein the basic is located between the plurality of basic block matrices at corresponding matrix positions.
- the connection between blocks may be a two-dimensional connection.
- Figure la is a plan view showing the basic blocks of the FPGA.
- Figure lb is a schematic diagram showing the topology of the switch matrix of the FPGA.
- FIG. 2 is a diagram showing an example of a logic block.
- FIG. 3 is a view showing an example of switches in the switch matrix.
- Fig. 4 is a diagram showing an FPGA realized by arranging basic blocks.
- Figure 5 is a plan view of the basic blocks of the 3D FPGA.
- FIG. 6a is a three-dimensional view of the switch matrix part of a three-dimensional FPGA.
- FIG. 6b is a schematic view showing the topology of the switch matrix of the 3D FPGA.
- FIG. 7 is a view showing a part of a three-dimensional FPGA in which the switch matrix of FIG. 6 a is arranged in a three-dimensional direction.
- Figure 8 shows an FPGA in which a 3D FPGA is embedded in a 2D chip layer by layer.
- Figure 9 shows a 3D FPGA embedded in a 2D chip layer by layer, with the end of the wiring in the z direction It is a figure showing connected FPGA.
- Fig. 10 is a diagram showing an FPGA in which a three-dimensional FPGA is embedded in a two-dimensional chip for each basic block adjacent in the z-axis direction.
- FIG. 11 is a diagram showing an FPGA in which a three-dimensional FPGA is embedded in a two-dimensional chip for each basic block adjacent in the z-axis direction and the end of the wiring in the z direction is connected.
- FIG. 12 is a diagram showing connections between basic blocks in an example of an FPGA in which four-dimensional FPGAs are embedded in a two-dimensional chip for each X-y plane.
- FIG. 13 is a schematic view of the topology of a four-dimensional switch matrix.
- FIG. 14 is a figure which shows a part of FPGA which connected the end of x, y axis between adjacent x-y planes of FPGA of FIG.
- FIG. 15 is a diagram showing an FPGA embedded in a two-dimensional chip by arranging basic blocks adjacent to each other in the z-axis direction in the vertical direction and arranging the three-dimensional FPGA in the two-dimensional chip.
- FIG. 16 is a diagram showing an FPGA embedded in a two-dimensional chip by arranging basic blocks adjacent in the z-axis direction in the three-dimensional FPGA in the vertical direction.
- FIG. 17a is a diagram showing a layout tile for realizing the embedding method of FIG.
- FIG. 17b is a diagram showing one region constituting the FPGA of FIG.
- FIG. 18a is a diagram showing a layout tile for realizing the embedding method of FIG.
- FIG. 18b is a diagram showing one region constituting the FPGA of FIG.
- FIG. 19a shows a layout tile for implementing the embedding method of FIG. 12.
- FIG. 19 b is a diagram showing one region constituting the FPGA of FIG.
- FIG. 20 is a diagram showing an FPGA configured by embedding a topology of a 5 dimensional or 6 dimensional switch in a 2 dimensional chip.
- Fig. 21a is a diagram showing a two-dimensional switch topology.
- FIG. 21b is a diagram showing a three-dimensional switch topology.
- FIG. 21c is a diagram showing a four-dimensional switch topology.
- FIG. 22 is a graph showing the required number of switches per basic block against the LUT density.
- FIG. 23 is a diagram showing an FPGA configured by a three-dimensional chip.
- Fig. 8 shows an F PGA 8 configured by embedding a three-dimensional FPGA with three rows and three columns and four layers into a two-dimensional chip.
- the first layer 801 to the fourth layer 804, each having rectangular areas congruent with each other, are two-dimensionally arranged clockwise.
- the first layer 801 to the fourth layer 804 are not arranged in layers because they are arranged in a two-dimensional manner. It is called “layer”.
- the wiring channel 506 connecting the layers is between the first layer 801 and the second layer 802, between the second layer 802 and the third layer 803, and between the third layer 803 and the fourth layer 804.
- each layer may be arranged in a two-dimensional plane and similarly connected. At that time, it is desirable to lay the layers in a folded manner so that the aspect ratio of the chip is reduced. Also, in the embedding method of FIG. 8, the first layer 801 and the fourth layer 804 are the end points of the wiring channel in the vertical direction of the three-dimensional FPGA, but in the FPGA 8, the first layer 801 and the fourth layer 804 Because they are adjacent to each other, they can be easily connected.
- each basic block 510 constituting the FPGA 8 has the same switch matrix 501 as the basic block shown in FIG. 5, a logic block 502 including variable logic elements and storage elements, a switch matrix 501, and a logic block 502 Wiring 505 that connects Have 503/504/506!
- a logic block 502 including variable logic elements and storage elements As the basic block shown in FIG. 5, a logic block 502 including variable logic elements and storage elements, a switch matrix 501, and a logic block 502 Wiring 505 that connects Have 503/504/506!
- it since it is connected between the first layer 801 to the fourth layer 804 arranged in a two-dimensional manner, it extends in the horizontal direction with respect to the chip because it is connected with the FPGA 8 and the self-wire line yanonere 506. There is.
- FIG. 9 shows that in the FPGA 8 in which a three-dimensional FPGA having three rows, three columns, and four layers is also embedded in a two-dimensional chip, the wiring channel 901 connects between the first layer 801 and the fourth layer 804.
- the FPGA 9 is shown configured in such a way that direction wiring loops. By doing this, the wiring area on the chip can be used more effectively than the structure of FIG.
- each basic block matrix can perform the function corresponding to the two-dimensional arrayed basic blocks provided in the conventional two-dimensional FPGA.
- the basic blocks located at corresponding matrix positions (in the same row and in the same column) among the plurality of basic block matrices are connected to each other.
- connections between basic blocks can be realized via a smaller number of switch matrices compared to conventional two-dimensional FPGAs, which are the same as conventional three-dimensional FPGAs.
- the logic density can be improved.
- the plurality of basic block matrices are arranged in a plane.
- the FPGA 8 '9 can solve the problems of the conventional two-dimensional and three-dimensional FPGAs.
- the basic blocks constituting each of the above basic block matrices are arranged in a group in a predetermined region (the first layer 801 to the fourth layer 804).
- connections between basic blocks located at corresponding matrix positions among the plurality of basic block matrices are one-dimensional connections.
- one-dimensional connections are added to two-dimensional connections in each basic block matrix.
- an FPGA 10 configured by an embedding method of arranging in a dimensional array is shown.
- an area 1010 is an arrangement of basic blocks existing in the upper right corner in each layer. That is, basic blocks 1001 to 1004 are two-dimensionally clockwise arranged in area 1010 in basic block 510 at the upper right corner in each of first layer 801 to fourth layer 804 in FPGA 8 in FIG. 8.
- FIG. 11 shows an FPGA 11 formed by connecting the basic block 1001 and the basic block 1004 by the wiring channel 1101 and looping the wiring channel in the vertical direction, as in the embedding method of FIG. Ru. By doing this, the wiring area on the chip can be used more effectively than the structure of FIG.
- the embedding method as shown in FIGS. 10 and 11 has an advantage that the wiring density is difficult to increase as compared with the embedding methods as shown in FIGS. 8 and 9. This advantage avoids the problem of increased chip area due to wiring.
- each basic block matrix is conventionally It can perform the function equivalent to the basic blocks arranged in a two-dimensional array that were provided in the two-dimensional FPGA.
- FPGAl it is possible to realize the connection between basic blocks via a smaller number of switch matrices, as compared to a conventional two-dimensional FPGA, which is the same as a conventional three-dimensional FPGA.
- the logic density can be improved.
- the plurality of basic block matrices are arranged in a plane. This eliminates the need to form fine vertical wires that are difficult to form in FPGPG10 ⁇ 11, which facilitates manufacture compared to conventional three-dimensional FPGAs, and also enables conventional three-dimensional F The problem of heat dissipation in PGA can also be suppressed.
- connections between basic blocks located at corresponding matrix positions among the plurality of basic block matrices are one-dimensional connections.
- basic blocks can be connected three-dimensionally by forming one-dimensional connections with respect to two-dimensional connections in each basic block matrix.
- FIG. 15 shows an FPGA 15 in which a three-dimensional FPGA having three rows, five columns, and four layers is embedded in a two-dimensional chip. Similar to the FPGA 10 of FIG. 10, the force of arranging the basic blocks in the same position in each layer together in one area is different from the FPGA 10, in the FPGA 15 the basic blocks 1001 to 1004 in the same position in each layer are in the area (z plane) 1510 It is arranged linearly in the vertical direction inside. In this way, the aspect ratio is increased and the mounting efficiency is degraded, but it is possible to reduce the aspect ratio by reducing the number of rows and the number of layers.
- FIG. 16 shows an FPGA 16 configured by connecting a region 1510 and a region 1520 by a wiring channel 1601 to the FPGA 15 of FIG. By doing this, the wiring area on the chip can be used more effectively than the structure of FIG.
- the top basic block in the area 1510 Focusing on the basic blocks at the top of the other regions, the noted basic blocks are arranged in a matrix and sequentially connected along the row direction and the column direction. It will be. This is the same even when focusing on the second top, the third top, and the bottom basic blocks in each region.
- each basic block matrix is It can perform the function equivalent to the basic block arranged in a two-dimensional array provided in the dimensional FPGA.
- the basic blocks located at corresponding matrix positions (in the same row and in the same column) among the plurality of basic block matrices are connected to each other.
- the connection between the basic blocks can be realized via a smaller number of switch matrices as compared to the conventional two-dimensional FPGA, which is the same as the conventional three-dimensional FPGA.
- the logic density can be improved.
- the plurality of basic block matrices are arranged in a plane. This eliminates the need to form fine vertical wires that are difficult to form in FPG 15.16, which facilitates manufacture as compared to conventional three-dimensional FPGAs, and also enables the conventional three-dimensional F The problem of heat dissipation in PGA can also be suppressed.
- the FPGA 15 '16 can solve the problems of the conventional two-dimensional and three-dimensional FPGAs.
- connections between basic blocks located at corresponding matrix positions among the plurality of basic block matrices are one-dimensional connections.
- basic blocks can be connected three-dimensionally by forming one-dimensional connections with respect to two-dimensional connections in each basic block matrix.
- Fig. 12 shows the FP constructed by embedding the topology of the 4D switch in the 2D chip. GA12 is shown.
- the FPGA 12 is composed of a basic block 1205 provided with a switch matrix (see FIG. 13 described later) having a four-dimensional topology. In the following, the four axes corresponding to four dimensions are denoted as x, y, u and v.
- the FPGA 12 in FIG. 12 is composed of 4 ⁇ 4 in the x ⁇ y direction and 3 ⁇ 3 in the uv direction, for a total of 144 basic blocks.
- the wiring channel 1201 is a y-direction
- the wiring channel 1202 is an x-direction
- the wiring channel 1203 is a v-direction
- the wiring channel 1204 is a u-direction wiring channel.
- the force and other switch matrixes showing only the wiring channels around one switch matrix are the same.
- Wiring channels in the x, y directions 1201 ⁇ 1202 connect the switch matrices in the x ⁇ y plane (region) 1206.
- Each XY plane is arranged in a two-dimensional plane to form an u-V plane.
- the wiring channels 12 03 ⁇ 1204 in the direction of the uv axis are connected between the adjacent XY planes in the uv plane. That is, the wiring channels 1203 and 1204 connect the same x ⁇ y positions (ie, the same row and the same column) of the switch matrix between the two x ⁇ y planes.
- FIG. 13 shows the topology of the four-dimensional switch matrix 13.
- FIG. 13 shows the connection from the wiring channel 1204a, that is, from the left side of the wiring channel in the u-axis direction, to the wiring channel 1201 '1201b' 1202a '1202b' 1203a '1203b' 1204b in each axial direction. .
- the arrows indicate the direction in which the connection is present, and there is a connection in each axial direction. In this case, only one direction of connection is shown. The same applies to both sides of the other axis.
- the four-dimensional switch matrix 13 there are switches connecting wiring channels in seven directions, so the number of switches per track increases compared to the two-dimensional or three-dimensional switch matrix. become.
- the increase in the number of interconnect channel tracks decreases, and as a result, the number of transistors can be reduced.
- FIGS. 9 and 11 the ends of the wiring channels in the z direction are connected in a loop.
- the ends of the wiring channels in the x and y directions can be connected to each other at the boundary of the adjacent xy planes on the uv plane.
- FIG. 14 shows the boundary of the xy plane 1 206 and the wiring channels in the x and y directions.
- the wiring channels 1401 connect the ends of the wiring channels in the X and y directions. In any embedding method, if the ends of the wiring channels in each axial direction are close to each other, they can be connected by connecting them. Wiring flexibility can be increased without increasing the number of switches.
- each basic block matrix can perform a function equivalent to a basic block arranged in a two-dimensional array provided in a conventional two-dimensional FPGA.
- the basic blocks located at corresponding matrix positions are connected to each other among the plurality of basic block matrices.
- connections between basic blocks can be realized via a smaller number of switch matrices, as compared to conventional two-dimensional FPGAs, which are the same as conventional three-dimensional FPGAs.
- the logic density can be improved.
- the plurality of basic block matrices are arranged in a plane.
- the FPGA 12 can solve the problems of the conventional two-dimensional and three-dimensional FPGAs.
- the basic blocks constituting each of the above basic block matrices are arranged in a group in a predetermined area (for example, the area 1206).
- connection between the basic blocks located at corresponding matrix positions among the plurality of basic block matrices is a two-dimensional connection in the u and V directions.
- FIG. 8 and FIG. 9 and FIG. 12 show basic blocks arranged in the xy plane direction (in FIG. 8 and FIG. 9, the basic blocks arranged inside of the first to fourth layers 801 to 804 respectively; 10, 11 and 15, and 16 are basic blocks arranged in the z-axis direction (each in FIGS. 10 and 11).
- This is a hierarchical structure in which the basic blocks arranged inside the area 1010, and the basic blocks arranged inside each area 1510 in FIGS. 15 and 16) are subregions.
- the uniqueness of the present invention will be described in terms of hierarchical structure.
- Patent Documents 4 and 10 the interior of the subregion is in a two-dimensional array, which looks like the appearance of FIG. 12 at first glance.
- the wiring channel of the super region exists in a lattice between the sub region and sub region, and the connection from the sub region to the wiring channel of the super region is limited to the edge of the sub region.
- This structure requires many switches to connect subregions and superregions, and again, it is not suitable for constructing large subregions.
- multi-layer wiring is currently used, there is no advantage in limiting the wiring of the super region to the edge of the sub region in this way.
- the switch matrix for each of the logical elements (basic blocks) inside the sub region which also serves as the switch matrix of the super region.
- the switch matrix of the super region Inside the subregion, there is a connection only between the switch matrixes of adjacent logic elements in each axial direction, and the wiring channel of the super region is also connected to each subregion. Since the connection between the switch matrices of the logic elements present at the same position in the cluster is connected and the subregions are more closely connected, the number of switches increases even if the size of the subregion is increased. Hateful.
- An embedding method may also be considered that can not be intentionally divided into subregions and superregions.
- Embedding an FPGA with a topology of five or more dimensions in two dimensions can be easily realized by applying three or four dimensional embedding.
- Embedding an odd-dimensional topology into a two-dimensional chip involves difficulties in layout.
- the embedding method of FIGS. 15 and 16 can easily realize the force that can be easily realized by arranging the same tile repeatedly.
- the method of embedding is somewhat difficult.
- a tile or layout tile is a unit rectangular area constituting an array, and is a layout of circuits constituting a basic block in a two-dimensional rectangular area.
- FIG. 17 a shows a basic block layout tile 1701 for implementing the embedding method of the FPGA 11 shown in FIG.
- FIG. 17b shows area 1010 of FIG. 10 formed by arranging layout tile 1701 and layout tiles 1702-1704 similar thereto.
- the embedding shown in FIG. 11 is realized by arranging the regions 1010 shown in FIG. 17b in a two-dimensional array.
- the layout tiles 1701 to 1704 in FIG. 17b are slightly different in the layout (layout) of the wiring of the wiring channels 503 and 504, respectively, and it is necessary to slightly change each design. As the array size is increased, more basic blocks with one-dimensional connections are folded and placed inside area 1010, which makes layout tile design more complicated.
- FIG. 18a shows a layout tile 1801 of basic blocks for realizing the embedding method of the FPGA 16 shown in FIG.
- FIG. 18 b shows the area 1510 of FIG. 16 formed by arranging four layout tiles 180 1.
- the embedding shown in FIG. 16 is realized by arranging the regions 1510 shown in FIG. 18b in a two-dimensional array.
- all tiles can be easily realized by arranging the same layout tile 1801. Ru.
- FIGS. 15 and 16 can be realized by arranging identical layout tiles, in order to reduce the aspect ratio, the number of basic blocks in the X direction is increased in the example of FIGS. 15 and 16, It is necessary to reduce the number of basic blocks in the y and z directions. However, in terms of routability, it is preferable that the number of basic blocks aligned in each axial direction be the same, that is, close to a cube.
- an embedding method intermediate to the embedding methods of FIGS. 15 and 16 and the embedding methods of FIGS. 10 and 11 can be considered.
- this method is a method of reducing the aspect ratio of the whole chip and creating an array close to a cube by folding the inside of the region 1510 a small number of times and simultaneously reducing the number of basic blocks in the y and z directions.
- the layout tiles are relatively easy to design because the number of folds is reduced.
- FIG. 19a shows a layout tile 1901 of basic blocks for realizing the method of embedding the four-dimensional topology of FIG. 12 in a two-dimensional chip.
- FIG. 19 b shows the area 1206 of FIG. 12 formed by arranging layout tile 1 901 4 ⁇ 4. Also, the embedding method of FIG. 12 is realized by arranging this area 1206 in 3 ⁇ 3.
- the layout tile 1901 of FIG. 19a is similar to a repeatable tile structure including mixed medium long lines found in Patent Documents 1, 2, 5, 8 and 9, but each of the above Patent Documents is a switch matrix. No mention is made of the internal topology (see Figure 13). Further, in the embodiments in the above-mentioned patent documents, there is no connection (1301 in FIG. 13) between the intermediate long wiring and the single long wiring in the return direction also in the case of! Deviation, which is different from the four-dimensional topology.
- the tile in the return direction through the middle long line can be physically connected at the shortest distance by passing through a single long line without passing through the middle long line, so the switch in the return direction is It is not necessary.
- the middle long line is made longer, it is preferable to connect to the switch matrix in the middle of the middle long line, thereby increasing the switch number.
- the wiring wire considered to correspond to the middle long line of each of the above-mentioned patent documents The channels 1203 1204 connect to the switch matrix only at their endpoints. Also, by connecting in the wiring matrix of each axial direction in the switch matrix, the wiring possibility is significantly improved, and the number of channel tracks can be significantly reduced. Therefore, the number of switches can be significantly reduced.
- the middle long lines found in Patent Documents 1, 2, 5, 8 and 9 (a wire having a length separating one or more basic blocks in the axial direction without being divided by a switch)
- wires of various lengths such as long wires (wires that have a length that penetrates the entire FPGA in the axial direction without being divided by switches) in each axial direction, or to use logic blocks in patent documents Hierarchical structures such as 6, 7 and 11 and patent documents 4 and 10 are also effective.
- Simulations were performed on an FPGA with a 2 to 4 dimensional topology using benchmark circuits and CAD tools, and the degree of integration and the required number of switches per basic block were examined.
- FIGS. 21a to 21c the switch topology of FIGS. 21a to 21c is assumed.
- Fig. 21a is two-dimensional (corresponding to Fig. La and Fig. Lb)
- Fig. 21b is three-dimensional (corresponding to Fig. 5, Fig. 6a and Fig. 6b)
- Fig. 21c is four-dimensional (corresponding to Fig. 12 and Fig. 13).
- the switch topology is shown, and the end points of the wiring in each axial direction are indicated by open circles, and the sides connecting the open circles indicate the presence of switches!
- FIG. 22 shows the required number of switches per basic block for the LUT density.
- the curves of the 2 to 4 dimensional topology show simulation results, and the curves of the 5 dimensional topology show those predicted from the results of 2 to 4 dimensions. From FIG. 22, it can be seen that the number of switches is minimized in two dimensions in several LUTs of LUT20, in three dimensions in 20 to 400, in four dimensions in 400 to 60000, and in five dimensions in 60000. Assuming that the switch is a pass transistor, 6 FPGAs with 2D topology, 15 in 3D, and 27 in 4D are required for each track, and the higher order elements The number of switches needed will increase.
- the number of wiring channel tracks required for an FPGA having a multidimensional topology is far greater than the number of wiring channel tracks required for an FPGA having a two-dimensional topology.
- the multi-dimensional topology is extremely advantageous in highly integrated FPGAs because it is reduced.
- Embedding a topology of five or more dimensions can be realized by applying a three-dimensional or four-dimensional embedding method.
- FIG. 20 shows an FPGA 20 configured by embedding a topology of a five- or six-dimensional switch in a two-dimensional chip.
- an area (subregion) 2001 in which basic blocks 2002 are arranged in a two-dimensional array is further arranged in a two-dimensional array, and between the basic blocks existing at the same position in the adjacent area 2001.
- the wiring channel is connected with the wiring channel in two directions of 2003 ⁇ 2004!
- the FPGA 20 has two axial directions corresponding to the wiring channel 2003 and 2004. It is an FPGA with a five-dimensional topology that has been pursued in terms of dimensions.
- the FPGA 20 has an FP having a six-dimensional topology in which two axial directions (dimensions) corresponding to the wiring channel 2 003 ⁇ 2004 are added. It becomes GA.
- FIG. 23 shows an FPGA 23 configured by a three-dimensional chip.
- a region (subregion) 2301 in which basic blocks 2302 are arranged in a two-dimensional array (subregions) 2301 is stacked in multiple layers, and one direction is between the basic blocks existing in the same position between adjacent subregions (layers).
- the wiring channel 2303 of If the area 2301 is internally connected in a three-dimensional topology as shown in FIGS. 8-11, 15 or 16, the FPGA 23 is four-dimensional following one axial direction (dimension) corresponding to the wiring channel 2303. It will be an FPGA with a topology. Further, if the inside of the area 2301 is connected in a four-dimensional topology as shown in FIG. 12, the FP GA 23 is an FPGA having a five-dimensional topological area in which one axial direction (dimension) corresponding to the wiring channel 2203 is added. It becomes.
- the FPGA in this embodiment is an FPGA having a multidimensional topology, and has a basic block configured by including a switch matrix (SM) having a multidimensional topology and a logic block (LB). These basic blocks are arranged in two or three dimensions and connected in a multidimensional topology.
- SM switch matrix
- LB logic block
- the present invention can be applied to other than FPGA. That is, the present invention is also applicable to a coarse-grain programmable device configured to include functional circuits such as adders, multipliers, and memories instead of the logic blocks in the above-mentioned FPGA.
- logic blocks are replaced with adders, multipliers, memories, etc., and they are connected in a multidimensional topology, and according to the method shown in FIGS. 8-12, 15, 16, 20, 23 It can be realized by embedding in a dimensional plane or three dimensional space.
- the present invention is also applicable to a semiconductor integrated circuit partially having an FPGA or a coarse grain size programmable device.
- a plurality of basic blocks and switching elements such as a switch matrix provided in each basic block are provided, and the basic blocks are connected via the respective switching elements. Can be applied to programmable devices.
- n is a natural number
- a switch has a structure for transmitting a signal from at least one of the constituent wires to at least one of the constituent wires of the kth bundle of wires.
- n 6 at the first feature point, and the first and second wiring bundles are at the top and bottom, respectively.
- (1 1) (1 > Connect to the switch matrix of the basic block which exists apart from the basic block of 1), and the third and fourth wiring bundles are (m-1) (m> 1) basic pro
- i, j is a natural number of 1 or more, connect (m x i) columns (IX (j ⁇ l) + 2k ⁇ 1) rows (where The fifth and sixth wiring bundles of the basic block located at l ⁇ k ⁇ (l + l) Z2 are connected to the switch matrix of the basic block adjacent to the left and upper direction, respectively, and (m 1) + 1)
- the fifth and sixth wiring bundles of the basic block located in the column (IX (j-1) + 2k) rows (where 1 ⁇ k ⁇ l / 2) are located in the right and upward direction respectively
- the fifth and sixth basic blocks are connected to the switch matrix of the adjacent basic block and located in (
- the fifth and sixth wiring bundles of the basic block located at 1) / 2) are connected to the switch matrixes of the basic blocks adjacent to the right and lower direction, respectively, and the fifth and fifth basic blocks at other positions are connected.
- the wiring bundle 6 has a structure connected to the switch matrix of the adjacent basic blocks in the left and right directions, respectively.
- n 8
- the first wiring bundle is upper
- the second wiring bundle is right
- the third wiring is The lower bundle and the fourth wire bundle are connected to the switch matrix of the adjacent basic block in the left direction
- the fifth and seventh wire bundles are (1 1) (1> 1) in the upper and lower directions, respectively.
- the sixth and eighth wiring bundles separate (m-1) (m> 1) basic blocks in the right and left directions, respectively. Have a structure connected to the switch matrix of the existing basic blocks.
- the first wiring bundle is on the top
- the second wiring bundle is on the right
- the third is The wiring bundle is connected to the switch matrix of the basic block adjacent in the lower direction and the fourth wiring bundle in the left direction
- the fifth and sixth wiring bundles are connected in the upper and lower directions (1 1) (1>) It has a structure connected to the switch matrix of the basic block which exists apart from the basic block of 1).
- Industrial Applicability The present invention can be applied to programmable devices represented by FPGAs.
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)
- Design And Manufacture Of Integrated Circuits (AREA)
- Logic Circuits (AREA)
Abstract
Description
Claims
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2006512926A JP4576538B2 (ja) | 2004-05-12 | 2005-03-28 | 多次元のスイッチトポロジーを有する集積回路 |
| US11/596,011 US7768314B2 (en) | 2004-05-12 | 2005-03-28 | Integrated circuit with multidimensional switch topology |
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2004170571 | 2004-05-12 | ||
| JP2004-170571 | 2004-05-12 | ||
| JP2004-281528 | 2004-09-28 | ||
| JP2004281528 | 2004-09-28 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2005109646A1 true WO2005109646A1 (ja) | 2005-11-17 |
Family
ID=35320522
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2005/005755 Ceased WO2005109646A1 (ja) | 2004-05-12 | 2005-03-28 | 多次元のスイッチトポロジーを有する集積回路 |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US7768314B2 (ja) |
| JP (1) | JP4576538B2 (ja) |
| TW (1) | TW200537805A (ja) |
| WO (1) | WO2005109646A1 (ja) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006107518A3 (en) * | 2005-04-01 | 2006-12-21 | Massachusetts Inst Technology | Methods and apparatus for 3-d fpga design |
| JP2007165589A (ja) * | 2005-12-14 | 2007-06-28 | Sony Corp | プログラムロジックデバイスおよび半導体パッケージ |
| JP2011527543A (ja) * | 2008-07-09 | 2011-10-27 | ユニベルシテ ピエール エ マリー キュリー(パリ シズエム) | プログラマブルゲートアレイ並びにそのようなアレイの相互接続スイッチ及び論理ユニット |
| JP5032996B2 (ja) * | 2005-11-28 | 2012-09-26 | 太陽誘電株式会社 | 半導体装置 |
| CN109428589A (zh) * | 2017-08-28 | 2019-03-05 | 中科亿海微电子科技(苏州)有限公司 | 通道结构非均匀分布三维fpga |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI386826B (zh) * | 2008-03-11 | 2013-02-21 | Rdc Semiconductor Co Ltd | 双端邏輯元件的方位決定方法 |
| JP5158195B2 (ja) * | 2008-06-06 | 2013-03-06 | 日本電気株式会社 | 回路設計システムおよび回路設計方法 |
| US8330489B2 (en) * | 2009-04-28 | 2012-12-11 | International Business Machines Corporation | Universal inter-layer interconnect for multi-layer semiconductor stacks |
| DE102010036927A1 (de) * | 2010-08-10 | 2012-02-16 | Technische Universität München | Schaltmatrixvorrichtung für elektrische Signale |
| US9148348B2 (en) * | 2011-10-31 | 2015-09-29 | Hewlett-Packard Development Company, L.P. | Generating network topologies |
| US9385922B2 (en) | 2011-10-31 | 2016-07-05 | Hewlett Packard Enterprise Development Lp | Generating EGFT network topologies |
| CH709741A1 (de) * | 2014-06-05 | 2015-12-15 | Swisstradingbox Ag | Börsenhandelsplattform. |
| CN108881877B (zh) * | 2017-05-12 | 2020-07-24 | 京东方科技集团股份有限公司 | 显示处理装置及其显示处理方法以及显示装置 |
| CN108595748B (zh) * | 2018-03-09 | 2022-08-09 | 电子科技大学 | 一种反熔丝fpga可编程逻辑阵列的三维拓扑结构 |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5748942A (en) * | 1995-06-07 | 1998-05-05 | Xilinx, Inc. | Efficient three-dimensional layout method for logic cell arrays |
Family Cites Families (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US34363A (en) * | 1862-02-11 | Improvement in machinery for cleaning cotton | ||
| USRE34363E (en) | 1984-03-12 | 1993-08-31 | Xilinx, Inc. | Configurable electrical circuit having configurable logic elements and configurable interconnects |
| US5396126A (en) * | 1993-02-19 | 1995-03-07 | At&T Corp. | FPGA with distributed switch matrix |
| US5455525A (en) * | 1993-12-06 | 1995-10-03 | Intelligent Logic Systems, Inc. | Hierarchically-structured programmable logic array and system for interconnecting logic elements in the logic array |
| EP0698312A1 (en) | 1994-02-15 | 1996-02-28 | Xilinx, Inc. | Tile based architecture for fpga |
| US5682107A (en) * | 1994-04-01 | 1997-10-28 | Xilinx, Inc. | FPGA architecture with repeatable tiles including routing matrices and logic matrices |
| US5689195A (en) * | 1995-05-17 | 1997-11-18 | Altera Corporation | Programmable logic array integrated circuit devices |
| KR960003103A (ko) | 1994-06-30 | 1996-01-26 | 윌리엄 이. 힐러 | 연합 헤테로젠니우스 필드 프로그래머블 게이트 어레이 논리 모듈 및 그 형성방법 |
| DE69521028T2 (de) * | 1995-03-29 | 2001-09-06 | Agilent Technologies, Inc. | Vortreiberschaltung zum rauscharmen Schalten hoher Ströme in einer Last |
| US5594363A (en) * | 1995-04-07 | 1997-01-14 | Zycad Corporation | Logic cell and routing architecture in a field programmable gate array |
| US5621342A (en) * | 1995-10-27 | 1997-04-15 | Philips Electronics North America Corporation | Low-power CMOS driver circuit capable of operating at high frequencies |
| US6094066A (en) * | 1996-08-03 | 2000-07-25 | Mission Research Corporation | Tiered routing architecture for field programmable gate arrays |
| US6300794B1 (en) * | 1996-10-10 | 2001-10-09 | Altera Corporation | Programmable logic device with hierarchical interconnection resources |
| US5883526A (en) * | 1997-04-17 | 1999-03-16 | Altera Corporation | Hierarchical interconnect for programmable logic devices |
| US5914616A (en) * | 1997-02-26 | 1999-06-22 | Xilinx, Inc. | FPGA repeatable interconnect structure with hierarchical interconnect lines |
| US5942913A (en) * | 1997-03-20 | 1999-08-24 | Xilinx, Inc. | FPGA repeatable interconnect structure with bidirectional and unidirectional interconnect lines |
| US5963050A (en) | 1997-02-26 | 1999-10-05 | Xilinx, Inc. | Configurable logic element with fast feedback paths |
| JPH10290147A (ja) * | 1997-04-14 | 1998-10-27 | Mitsubishi Electric Corp | 遅延量可変回路 |
| US6084429A (en) * | 1998-04-24 | 2000-07-04 | Xilinx, Inc. | PLD having a window pane architecture with segmented and staggered interconnect wiring between logic block arrays |
| US6278290B1 (en) * | 1999-08-13 | 2001-08-21 | Xilinx, Inc. | Method and circuit for operating programmable logic devices during power-up and stand-by modes |
| JP2001144603A (ja) * | 1999-11-18 | 2001-05-25 | Oki Micro Design Co Ltd | レベルシフタ回路およびそれを含むデータ出力回路 |
| US6262599B1 (en) * | 2000-04-06 | 2001-07-17 | International Business Machines Corporation | Level shifting CMOS I/O buffer |
| JP4963144B2 (ja) * | 2000-06-22 | 2012-06-27 | ルネサスエレクトロニクス株式会社 | 半導体集積回路 |
| IT1318794B1 (it) * | 2000-08-29 | 2003-09-10 | St Microelectronics Srl | Circuito per il pilotaggio di un interruttore di potenza. |
| US6815276B2 (en) * | 2002-10-03 | 2004-11-09 | Texas Instruments Incorporated | Segmented power MOSFET of safe operation |
| US7457303B2 (en) * | 2003-06-06 | 2008-11-25 | International Business Machines Corporation | One-bounce network |
| US7486619B2 (en) * | 2004-03-04 | 2009-02-03 | International Business Machines Corporation | Multidimensional switch network |
| US7215146B2 (en) * | 2004-10-29 | 2007-05-08 | Intel Corporation | High speed buffered level-up shifters |
| US20060225020A1 (en) * | 2005-04-01 | 2006-10-05 | Anantha Chandrakasan | Methods and apparatus for 3-D FPGA design |
| EP1801975A1 (en) * | 2005-12-21 | 2007-06-27 | STMicroelectronics S.r.l. | Output buffer |
| US20070252638A1 (en) * | 2006-04-26 | 2007-11-01 | Farrukh Aquil | Method and apparatus for temperature compensating off chip driver (OCD) circuit |
-
2005
- 2005-03-28 US US11/596,011 patent/US7768314B2/en not_active Expired - Fee Related
- 2005-03-28 WO PCT/JP2005/005755 patent/WO2005109646A1/ja not_active Ceased
- 2005-03-28 JP JP2006512926A patent/JP4576538B2/ja not_active Expired - Lifetime
- 2005-03-31 TW TW094110400A patent/TW200537805A/zh unknown
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5748942A (en) * | 1995-06-07 | 1998-05-05 | Xilinx, Inc. | Efficient three-dimensional layout method for logic cell arrays |
Non-Patent Citations (1)
| Title |
|---|
| ALEXANDER M.J. ET AL: "Three-dimensional field-programmable gate arrays, ASIC Conference and Exhibit.", PROCEEDINGS OF THE EIGHTH ANNUAL IEEE INTERNATIONAL., 22 September 1995 (1995-09-22), pages 253 - 256, XP010218016 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006107518A3 (en) * | 2005-04-01 | 2006-12-21 | Massachusetts Inst Technology | Methods and apparatus for 3-d fpga design |
| JP5032996B2 (ja) * | 2005-11-28 | 2012-09-26 | 太陽誘電株式会社 | 半導体装置 |
| JP2007165589A (ja) * | 2005-12-14 | 2007-06-28 | Sony Corp | プログラムロジックデバイスおよび半導体パッケージ |
| JP2011527543A (ja) * | 2008-07-09 | 2011-10-27 | ユニベルシテ ピエール エ マリー キュリー(パリ シズエム) | プログラマブルゲートアレイ並びにそのようなアレイの相互接続スイッチ及び論理ユニット |
| CN109428589A (zh) * | 2017-08-28 | 2019-03-05 | 中科亿海微电子科技(苏州)有限公司 | 通道结构非均匀分布三维fpga |
| CN109428589B (zh) * | 2017-08-28 | 2023-09-22 | 中科亿海微电子科技(苏州)有限公司 | 通道结构非均匀分布3d fpga |
Also Published As
| Publication number | Publication date |
|---|---|
| TW200537805A (en) | 2005-11-16 |
| JP4576538B2 (ja) | 2010-11-10 |
| US7768314B2 (en) | 2010-08-03 |
| US20090009215A1 (en) | 2009-01-08 |
| JPWO2005109646A1 (ja) | 2008-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2005109646A1 (ja) | 多次元のスイッチトポロジーを有する集積回路 | |
| US6275064B1 (en) | Symmetrical, extended and fast direct connections between variable grain blocks in FPGA integrated circuits | |
| US6150842A (en) | Variable grain architecture for FPGA integrated circuits | |
| US6526555B1 (en) | Method for layout and manufacture of gridless non manhattan semiconductor integrated circuits using compaction | |
| Farooq et al. | Tree-based heterogeneous FPGA architectures: application specific exploration and optimization | |
| US7295037B2 (en) | Configurable IC with routing circuits with offset connections | |
| US20070285124A1 (en) | Embedding Memory Between Tile Arrangement of a Configurable IC | |
| WO2000035093A1 (en) | Efficient interconnect network for use in fpga device having variable grain architecture | |
| Gaillardon et al. | A novel FPGA architecture based on ultrafine grain reconfigurable logic cells | |
| Lai et al. | Hierarchical interconnection structures for field programmable gate arrays | |
| US9257986B2 (en) | Rescaling | |
| US6034544A (en) | Programmable input/output block (IOB) in FPGA integrated circuits | |
| Padalia et al. | Automatic transistor and physical design of FPGA tiles from an architectural specification | |
| Siozios et al. | A novel framework for exploring 3-d fpgas with heterogeneous interconnect fabric | |
| JP5305806B2 (ja) | 3次元集積回路の設計方法及び3次元集積回路の設計プログラム | |
| Chiricescu et al. | Design and analysis of a dynamically reconfigurable three-dimensional FPGA | |
| US8788987B2 (en) | Rescaling | |
| JP5024530B2 (ja) | 三次元集積電気回路の配線構造及びそのレイアウト方法 | |
| Cho et al. | M/sup 2/R: Multilayer routing algorithm for high-performance MCMs | |
| US5982193A (en) | Input/output block (IOB) connections to MaxL lines, nor lines and dendrites in FPGA integrated circuits | |
| US6107823A (en) | Programmable control multiplexing for input/output blocks (IOBs) in FPGA integrated circuits | |
| CN1862968B (zh) | 集成电路以及设计集成电路的方法 | |
| Ran et al. | Designing a via-configurable regular fabric | |
| WO2008042109A2 (en) | Structure cluster and method in programmable logic circuit | |
| Ran et al. | An integrated design flow for a via-configurable gate array |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| WWE | Wipo information: entry into national phase |
Ref document number: 2006512926 Country of ref document: JP |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| WWW | Wipo information: withdrawn in national office |
Country of ref document: DE |
|
| 122 | Ep: pct application non-entry in european phase | ||
| WWE | Wipo information: entry into national phase |
Ref document number: 11596011 Country of ref document: US |