Interconnect structure for a programmable logic device

专利摘要:
The invention provides an FPGA interconnect structure preferably included in an array of identical tiles. A combination of single-length lines connecting to adjacent tiles and intermediate-length lines connecting to tiles several tiles away creates an interconnect hierarchy which allows any logic block to be connected to any other logic block, yet also allows for fast paths to both adjacent tiles and tiles some distance away. Longer interconnect lines may be included as a third level of hierarchy to permit interconnection of widely separated tiles. In a preferred embodiment, from a given tile an intermediate-length line connects to the tile three tiles away, then continues and connects to the tile six tiles away. In this embodiment, the intermediate-length line does not connect to the intervening tiles one, two, four, and five tiles away.
公开号:US20010007428A1
申请号:US09/759,051
申请日:2001-01-11
公开日:2001-07-12
发明作者:Steven Young；Kamal Chaudhary；Trevor Bauer
申请人:Young Steven P.；Kamal Chaudhary；Bauer Trevor J.；
IPC主号:H03K19-1737

专利说明:
[0001] 1. Field of the Invention [0001]
[0002] The invention relates to programmable integrated circuit devices, more particularly to the interconnect structure in a field programmable logic device. [0002]
[0003] 2. Description of the Background Art [0003]
[0004] Field programmable gate arrays (FPGAs) include logic blocks connectable through a programmable interconnect structure. The interconnect structure typically provides for connecting each logic block to each other logic block. Early FPGAs accomplished this by providing short interconnect segments that could be joined to each other and to input and output terminals of the logic blocks at programmable interconnection points (PIPS). AS these FPGAs become larger and more complex, the interconnect structure must also become both larger and more complex. In order to improve speed (performance), direct connections to adjacent logic blocks have been provided, and for transmitting a signal the distance of many logic blocks, longer lines have been provided. In order to save silicon area, less frequent PIPs have been provided. With fewer PIPs present, the routing is less flexible (for the same number of routing lines), but typically faster due to reduced loading. By removing only those PIPs which are least often used, routing flexibility can be minimally affected. Thus, there is a trade-off between performance, silicon area, number of routing lines, and routing flexibility. [0004]
[0005] Several U.S. Patents show such structures for interconnecting logic blocks in FPGAs. Freeman in U.S. Reissue Pat. No. Re 34,363 describes the first FPGA interconnect structure, and includes short routing segments and flexible connections as well as global lines for signals such as clock signals. Carter in U.S. Pat. No. 4,642,487 shows the addition of direct connections between adjacent logic blocks to the interconnect structure of Freeman. These direct connections provide fast paths between adjacent logic blocks. Greene et al in U.S. Pat. No. 5,073,729 shows a segmented interconnect structure with routing lines of varied lengths. Kean in U.S. Pat. No. 5,469,003 shows a hierarchical interconnect structure having lines of a short length connectable at boundaries to lines of a longer length extending between the boundaries, and larger boundaries with lines of even longer length extending between those boundaries. Kean shows in particular lines the length of one logic block connecting each logic block to the next, lines the length of four logic blocks connectable to each logic block they pass, and lines the length of sixteen logic blocks connectable at the length-four boundaries to the length-four lines but not connectable directly to the logic blocks. In Kean's architecture, adjacent logic blocks in two different hierarchical blocks (i.e., on either side of the boundaries) connect to each other differently than adjacent logic blocks in the same hierarchical block. [0005]
[0006] Pierce et al in U.S. Pat. No. 5,581,199 shows a tile-based interconnect structure with lines of varying lengths in which each tile in a rectangular array may be identical to each other tile. In the Pierce et al architecture, an interconnect line is part of the output structure of a logic block. Output lines of more than one length extend past other logic block input lines to which the logic block output lines can be connected. All of the above-referenced patents are incorporated herein by reference, and can be reviewed for more understanding of prior art routing structures in FPGAs. [0006]
[0007] In the interconnect structures described by Freeman and Greene et al, each path is formed by traversing a series of programmably concatenated interconnect lines, i.e., a series of relatively short interconnect lines are programmably connected end to end to form a longer path. The relatively large number of programmable connections on a given signal path introduces delay into the signal path and therefore reduces the performance of the FPGA. Such interconnect structures are called “general interconnect”. [0007]
[0008] The direct connections first described by Carter and included in the architecture of Kean provide fast paths between adjacent logic blocks, but in Carter's structure general interconnect must still be used to traverse the distance between any two blocks that are not adjacent. Therefore, circuits large enough or complex enough to require interconnecting signals between non-adjacent blocks (which frequently occur) must use the general interconnect to make these connections. For short paths, general interconnect is slower than direct interconnect, because general interconnect must be connected through several PIPs, or, if long lines are used, must be buffered to accommodate long or heavily loaded signals, introducing delay. Additionally, it is inefficient in terms of silicon area to use long lines for short paths that may be traversing only a few logic blocks, since the long lines can otherwise be used for longer paths. Further, since software that implements a logic design in an FPGA typically places interconnected logic in close proximity, structures that take advantage of this placement strategy will work well with the software, resulting in shorter compilation times for routing software and more efficient circuit implementations. [0008]
[0009] Interconnect lines called “quad lines” are included in the XC4000EX FPGAs from Xilinx, Inc., and described on pages 4-32 through 4-37 of the Xilinx 1996 Data Book entitled “The Programmable Logic Data Book”, available from Xilinx, Inc., 2100 Logic Drive, San Jose, Calif. 95124, which are incorporated herein by reference. However, since each quad line contacts every tile that it traverses, these lines have a large number of PIPs, each of which adds RC delay. [0009]
[0010] Pierce et al provides fast paths between both adjacent logic blocks and logic blocks several tiles apart. The output lines of the Pierce et al architecture can each drive the inputs of a limited set of other logic blocks. However, the possible destinations are limited to selected logic blocks, and the interconnect lines can only access certain specific inputs of the destination logic blocks. [0010]
[0011] In each of the prior art structures recited above, each interconnect line has programmable connections to the inputs of other logic blocks. However, in the structures of Freeman, Carter, and Pierce et al, a given logic block input can be driven from either horizontal interconnect lines, or vertical interconnect lines, but not both. An alternative approach is to separate the interconnect lines from the logic block inputs by way of a routing matrix, which gives each interconnect line more flexible access to the logic block inputs. Such an architecture is described in U.S. Pat. No. 5,682,107 entitled “FPGA Architecture With Repeatable Tiles Including Routing Matrices and Logic Matrices” by Tavana et al, issued Oct. 28, 1997, which is referenced above and incorporated herein by reference. In the structure of Tavana et al, most interconnect lines entering the tile connect to a routing matrix within the tile, rather than directly to logic block inputs or outputs. Connections between pairs of interconnect lines and between interconnect lines and logic block inputs are made through lines called “tile interconnect lines” that do not leave the tile. The advantage of having an extra interconnect line in a path from the edge of a tile to the logic block in the tile is that the routing matrix is flexible but consumes a relatively small amount of silicon area. A combination of PIPs can allow access from any line entering the tile to any desired input of a destination logic block. Yet the total number of PIPs is smaller than in many other interconnect structures. The disadvantage is that getting on and off the tile interconnect lines inserts a certain amount of delay into the path for each tile traversed. This delay inhibits the fast propagation of signals through the FPGA. Tavana et al have therefore provided long lines connectable to every tile they pass and double-length lines that bypass the tile interconnect lines in one tile. These lines can be used for signals that are traversing one or more tiles without accessing the logic blocks in the traversed tiles. [0011]
[0012] Kean separates the interconnect lines from the logic block inputs using input multiplexer switches, which provide routing flexibility to the inputs. [0012]
[0013] Since the slowest signal path between logic blocks typically determines the performance of a circuit, it is advantageous to make the slowest path as fast as possible. One way to accomplish this is to design the interconnect structure such that there is a relatively uniform delay on all signal paths throughout an FPGA. In the above routing structures, a typical distribution of delays on signal paths shows a few signal paths with significantly greater delay than the average. These signal paths are typically those with large “RC trees”, i.e., signal paths which traverse a resistor (such as an unbuffered PIP), then have a large capacitance on the destination side of the resistor. An interconnect structure with relatively uniform delay could be better realized if large capacitances on a signal path (e.g., longer interconnect lines) were predictably placed on the source side of the resistor, or as close as possible to the source end of the signal path. [0013]
[0014] High fanout signals have large capacitance and are often slower than low fanout signals. Prior art routing structures had high-fanout signal routing with relatively large RC delay. An interconnect structure should ideally provide high-fanout signal routing with a delay comparable to that of other signals. [0014]
[0015] It is therefore desirable to find an interconnect structure that allows: 1) uniformly fast propagation of signals, including high-fanout signals, throughout the FPGA; 2) implementation of localized circuits in non-adjacent as well as adjacent blocks using fast paths; 3) ease of use by software; 4) efficient implementation of commonly used logic functions; and 5) a high degree of routing flexibility per silicon area consumed. [0015] SUMMARY OF THE INVENTION
[0016] According to the invention, an FPGA interconnect structure includes a combination of wiring segment lengths and connections to logic blocks such that a given logic block can be connected through fast paths both to adjacent logic blocks and to logic blocks several tiles away. In the preferred mode, the FPGA includes a two-dimensional array of identical tiles. In each tile is a logic block. Also in each tile are programmable interconnection points (PIPs) and segments of interconnect lines that adjoin segments of interconnect lines in adjacent tiles. The adjoined segments form interconnect lines extending through several tiles to PIPs in other tiles some distance away. A combination of lines connecting to adjacent tiles (called single-length lines) and lines at least three tiles long connecting a first tile to at least second and third tiles at least two and three tiles away (called intermediate-length lines) creates an interconnect hierarchy which allows any logic block to be connected to any other logic block, and yet also allows for fast paths both to adjacent tiles and to tiles some distance away. Longer interconnect lines (called long lines) may be included as a third level of hierarchy to permit efficient interconnection of widely separated tiles. Long lines can span the entire width of the tile array, or can be programmably segmented into two or more shorter long lines. In one embodiment, long lines are distinguished from intermediate-length lines in that a pattern of PIPs spanning two or more tiles is repeated along the length of the long line. When the size of the tile array is increased, more instances of the pattern occur. By contrast, intermediate-length lines according to the invention are of a predetermined length that does not alter when the size of the array is increased. When the size of the tile array is increased, additional intermediate-length lines are added. [0016]
[0017] A unique aspect of the invention is having an interconnect line (specifically an intermediate-length line or a long line) that programmably connects to logic blocks in at least three separate tiles, while extending through at least one tile in which no PIPs connect to the interconnect line. Such an interconnect line is said to be “unconnectable” to the tile in which no PIPs connect to the interconnect line and also “unconnectable” to the logic block in the tile, although connections can be made from the interconnect line to the logic block through PIPs in other tiles. A programmable connection from an interconnect line to a logic block in a given tile may be made: a) directly to a logic block input; b) through one or more PIPs connected to the interconnect line in the same tile; or c) through one or more PIPs connected to the interconnect line in the same tile and through one or more single-length lines. If any such programmable connection from an interconnect line to a logic block in a given tile can be made, the interconnect line is said to be “connectable” to the logic block and to the tile. [0017]
[0018] In a preferred embodiment, from an originating tile an intermediate-length line connects to the tile three tiles away (i.e., separated by two tiles from the originating tile), then continues and connects to the tile six tiles away (i.e., separated by five tiles from the originating tile). This intermediate-length line (called a “hex line”) does not connect to the intervening tiles one, two, four, and five tiles away. Instead, these tiles are reached indirectly by using single-length lines from the originating tile, the third tile, or the sixth tile. Connecting to only three of the seven tiles traversed by the hex line reduces the number of PIPs, and therefore reduces the silicon area required to form the interconnect line (thereby lowering the cost of the FPGA) and also reduces the capacitance added by PIPs (thereby increasing FPGA performance). Yet the combination of this intermediate-length routing, direct connections to neighboring tiles, and long lines to distant tiles allows highly flexible routing of signals. [0018]
[0019] Many modifications of the interconnect structure of the invention are possible. One such modification is the use of asymmetrical PIP patterns on interconnect lines (intermediate-length lines or long lines) such that the interconnect line connects three logic blocks, with two of the three logic blocks being in adjacent tiles, while the third logic block is in a tile separated from the other two tiles. [0019]
[0020] As FPGAs grow larger, the amount of routing required per tile grows larger. Therefore, the silicon area required to implement the necessary PIPs tends to grow larger, and the silicon area per tile increases as the number of tiles in an FPGA increases. It is desirable to reduce the number of PIPs required per tile. [0020]
[0021] In one embodiment, only the single-length lines have connections to the logic block inputs. The intermediate-length lines have connections to each other and to single-length lines, but not to the logic block inputs. The long lines have connections to the intermediate-length lines, but not to the single-length lines or to the logic block inputs. Therefore, the number of PIPs in a tile is reduced. As with the intermediate-length lines, reducing the number of PIPs on a long line reduces both silicon area and capacitance on the long line, thereby reducing RC delay. In some embodiments, special tiles include additional connections between the various interconnect lines to facilitate distribution of high-fanout signals. In some embodiments, global lines having direct access to the logic block inputs are available for global signals such as clocks or other high fanout control signals. In some embodiments, the long lines have connections to each other in the repeatable tile instead of in special tiles. [0021]
[0022] In some embodiments, the single-length lines are driven by unbuffered PIPs and the intermediate-length and long lines are driven by buffered PIPs. Signals are typically routed on longer buffered interconnect lines first, then fanned out on shorter unbuffered lines. In this manner, large unbuffered RC trees are avoided, making delays on signal paths throughout the FPGA more uniform and improving performance. More uniform delays also make it easier to predict the performance of a circuit earlier in the design cycle. [0022] BRIEF DESCRIPTION OF THE DRAWINGS
[0023] FIG. 1 shows some of the single-length and intermediate-length interconnect lines according to one embodiment of the invention. FIG. 1 is a simplified representation of the embodiment of FIG. 2. [0023]
[0024] FIG. 2 shows a bus representation of a single tile in which intermediate-length lines are connected to every third and sixth tile (i.e., hex buses are used). [0024]
[0025] FIG. 2A is a slightly modified version of FIG. 2 showing a different representation of the single-length buses and switching structure [0025] 403.
[0026] FIG. 3 shows a detailed representation of a single tile in the preferred embodiment. All features shown in FIG. 2 are represented, and some additional features are included. [0026]
[0027] FIG. 4 shows the region of FIG. 3 labeled INTERCONNECT. [0027]
[0028] FIG. 5 shows the output multiplexer (OMUX) of FIG. 3. [0028]
[0029] FIG. 6 shows the input and output signals for the configurable logic element (CLE) of FIG. 3. [0029]
[0030] FIGS. 6A and 6B show the internal logic for the CLE of FIG. 3. To view the entire CLE, FIGS. 6A and 6B must be viewed together as shown in the key at the top of FIG. 6A. [0030]
[0031] FIG. 7 shows the input multiplexer (IMUX) of FIG. 3. [0031]
[0032] FIG. 8 shows a routing example of a signal path using hex and single-length lines. [0032]
[0033] FIG. 9 shows a routing example of a signal path using long lines, hex lines, and single-length lines. [0033]
[0034] FIG. 10 shows a routing example of a high fanout control signal delivered to a column of tiles. [0034]
[0035] FIG. 11 shows a routing example of a high fanout function generator input signal delivered to a column of tiles. [0035]
[0036] FIG. 12 shows an example of how to combine the interconnect of FIGS. [0036] 10 and/or 11 to deliver a high-fanout signal to an array of tiles. DETAILED DESCRIPTION OF THE DRAWINGS
[0037] The following drawing conventions are used throughout the figures. A small solid black dot at the intersection of two lines indicates a permanent electrical connection between the crossing lines. An open circle enclosing an intersection between two lines, or separating a single line into two line segments, indicates a programmable bidirectional connection between the lines or line segments (for example, a pass transistor which is turned on to make the connection). An open triangle at an intersection of two lines indicates a programmable connection with signal flow going onto the line pointed to by the apex of the triangle. (The signal is of course then present on the full length of the line. Thus, a triangle pointing in the opposite direction would have the same signal flow because the triangle points to the same line.) A triangle showing a connection between two buses indicates that at least one connection exists between the lines making up the buses. In accordance with one embodiment of the present invention, programmable connections are provided by using programmable interconnection points (PIPs), wherein each PIP includes at least one transistor. [0037]
[0038] A line which ends within a tile or block structure (i.e., does not extend to the border of a tile or block) is physically terminated within the tile or block. A line which extends to the border of the tile connects to a line on the next tile, which it contacts when two tiles are abutted together. Note that some lines which extend to an edge of a tile and thus into an adjacent tile change names at the tile boundary. [0038] Simplified Representation of the Inventive Concept
[0039] FIG. 1 shows in simplified form some of the single-length and intermediate-length interconnect lines according to one embodiment of the invention. FIG. 1 illustrates part of an array of tiles in an FPGA. In each tile is logic block L and a routing or switching structure R. Also in each tile are line segments which connect to line segments in adjacent tiles to form interconnect lines. Most of the interconnect lines in the FPGA are not shown in FIG. 1 so as not to obscure the connectivity of the invention. Three of the tiles, labeled TILE[0039] 1, TILE2, and TILE3 are shown in more detail, and short- and intermediate-length interconnect lines extending from TILE1, TILE2, and TILE3 are shown. (However, intermediate-length lines extending horizontally are not shown.) Single-length lines N, S, E, and W extend north, south, east, and west, respectively, from TILE2 to the four neighboring tiles. Equivalent lines not labeled extend from TILE1 and TILE3 as well. Identical structures are also present in all other tiles but are not shown in FIG. 1. These single-length lines N, S, E, and W extend between switching structures R in neighboring tiles; and have programmable connections to logic block L.
[0040] Intermediate-length lines also extend north, south, east and west from each tile to connect to other tiles three and six tiles away. Therefore in this embodiment the intermediate-length lines are “hex lines” as previously defined. Only the vertically extending (north and south) hex lines are shown, to avoid obscuring the connections. (Later figures show the complete connection pattern.) [0040]
[0041] From TILE[0041] 2, hex line 6VN extends to the north a length of six tiles, connecting to TILE1 three tiles to the north and to another tile (not shown) six tiles to the north. Also from TILE2, hex line 6VS extends to the south six tiles, connecting to TILE3 and to another tile (not shown) six tiles to the south. Hex line 6VM connects tile TILE2 to tiles TILE1 and TILE3. Thus, three hex lines extend vertically and connect to TILE2. Also in TILE2, line 6VS is programmably connectable to line 6VN, as indicated by the PIP joining the ends of these two lines. In this manner an easy connection is available to a tile twelve tiles away from the starting tile. Only one additional connection is needed to extend the connection to another interconnect line six tiles long.
[0042] Note that within TILE[0042] 2, only the single-length lines N, S, E, and W have access to logic block L. Thus a signal on a hex line must transfer to a single-length line to access logic block L in the destination tile. A signal entering TILE2 on line 6VS can connect to lines W, N, or E and thereby to logic block L in TILE2. Further, the signal now on single-length lines W, N or E can also access the logic block L in the tiles to the west, north or east of TILE2. In this embodiment, no connection from hex line 6VS to single-length line S is provided, as eliminating this connection saves silicon area and the tile to the south of TILE2 can be accessed through another path. In the embodiment of FIG. 1, a signal in TILE3 can reach the logic block of the tile to the south of TILE2 either by: 1) traversing two single-length lines to the north; or 2) taking hex line 6VM to TILE2, connecting to single-length line W or E, then entering switching structure R in TILE2 and “bouncing” (moving onto and off of an interconnect line in the same tile) from single-length line W or E to single-length line S inside the switching structure. The second alternative has one more PIP in the signal path than the first alternative, the PIP in switching structure R. In another embodiment, a connection to the south can of course be provided. Repeatable Tile Representation
[0043] FIG. 2 more completely illustrates a single tile. FIG. 2 shows bus segments of various widths, as indicated by diagonal slash lines having a number indicating the bus width. The number of lines in each bus segment can of course vary in different embodiments. In the embodiment of FIG. 2, the tiles are identical. Since the illustration is of a single tile, complete interconnect lines must be understood by realizing that bus segments extending to the top of FIG. 2 will contact bus segments at the bottom of another copy of FIG. 2 placed adjacent to the north. Similarly, bus segments extending to the right of FIG. 2 will contact bus segments of another copy of FIG. 2 placed adjacent to the right, and so forth. An array such as present in an FPGA is represented by placing many copies of FIG. 2 together. [0043] Hex Buses
[0044] In FIG. 2, 12-segment bus [0044] 6N extends horizontally through programmable region PROG-2, then turns north to become bus segment 47. Bus segment 47 extends to the top edge of the tile. Another copy of FIG. 2 placed adjacent and to the north of FIG. 2 will have a bus segment 46 extending to the bottom of the copy that contacts bus segment 47. Bus segment 46 extends vertically through the tile, but is offset right by one bus before it reaches the top of the tile, such that bus segment 46 in a given tile contacts bus segment 45 in the tile adjacent to the north. Similarly, at the tile boundary, bus segment 45 contacts bus segment 44, bus segment 44 contacts bus segment 43, bus segment 43 contacts bus segment 42, and bus segment 42 contacts bus segment 41. This offset arrangement results in a hex length bus extending through six adjacent tiles. Bus segment 41 extends into the tile from the bottom, then turns west and extends horizontally through programmable region PROG-2 as bus segment 6S. Bus segment 6V also extends horizontally through programmable region PROG-2, and connects to bus segment 44. Each tile therefore includes three bus segments 6N, 6V, and 6S extending into programmable region PROG-2. These bus segments enable the connections from a given tile to tiles three and six tiles away, which makes these intermediate-length lines hex lines as defined above. Bus segments 41, 42, 43, 44, 45, 46 and 47 together form a 72-line vertical “routing track” designated 6VN at the north end and 6VS at the south end of FIG. 2. Bus segments 42, 43, 44, 45, and 46 are called “interior segments”, and bus segments 41 and 47 are called “end segments”.
[0045] Horizontal 12-segment hex buses are similarly provided, each tile having portions [0045] 6W, 6H, and 6E extending into programmable region PROG-2. Bus segment 6E extends from programmable region PROG-2, then turns east and extends as bus segment 37 to the east edge of the tile. Bus segment 37 of a given tile contacts bus segment 36 in the adjacent identical tile to the east. Similarly to bus segments 41-47, bus segments 31-37 provide a hex length bus extending through six adjacent tiles. Bus segment 31 extends into programmable region PROG-2 as bus segment 6W. Bus segment 34 extends into programmable region PROG-2 as bus segment 6H. Bus segments 31, 32, 33, 34, 35, 36 and 37 together form a 72-line horizontal routing track designated 6HW at the west end and 6HE at the east end of FIG. 2.
[0046] Note that 72-line hex buses [0046] 6VN, 6VS, 6HW, 6HE are not typical buses in that the hex lines making up the buses do not all start in the same tile and end in the same tile, i.e., there is no fixed boundary for the bus. Twelve hex lines end, and another twelve hex lines begin, in each tile. This feature, among others, distinguishes the inventive structure over that of Kean and other hierarchical interconnect structures. In the hierarchical interconnect structures, an FPGA is divided hierarchically with periodic routing structures at the hierarchical boundaries. The present invention uses identical tiles, each of which tiles includes end segments of some of the lines. Other Buses
[0047] North, south, east, and west single-length buses (N, S, E, and W respectively), each 24 lines wide, are also provided. These single-length buses connect to each other within switching structure [0047] 403 located in programmable region PROG-2. In FIG. 2, single-length buses S, W and E are shown entering switching structure 403 in two positions, in order to correspond to FIG. 4. However, the bus segments labeled S, W, and E pass continuously through switching structure 403. The way these bus segments pass through switching structure 403 is clarified in FIG. 2A, which shows a different representation of switching structure 403, but is otherwise the same as FIG. 2.
[0048] In this embodiment, each single-length line can be connected to the input multiplexer IMUX-[0048] 2. Each single-length line can also be connected to some of the hex lines.
[0049] FIG. 2 includes another feature: horizontal and vertical long line buses. Each of the two long line buses is [0049] 12 lines wide. Two-line buses 437P, 438 (corresponding to lines LV0 and LV6, LH0 and LH6, respectively, in FIGS. 4 and 7) taken from each long line bus extend into programmable region PROG-2 where they can be programmably connected to drive hex lines. The two accessed long lines in each long line bus are separated from each other by six lines, and each long line is accessed every sixth tile. Two-line buses 437I, 438 taken from each long line bus extend from input multiplexer IMUX-2, which can provide signals to the long lines. (Two-line bus 437I includes the same two signals as two-line bus 437P.) The horizontal long line bus is designated LHW at the west end and LHE at the east end of FIG. 2. LHW and LHE extend through an offset region 422 (shown in FIG. 3) which offsets the bus by one line per tile. In this manner, each identical tile accesses two different lines of horizontal long line bus LHW/LHE, and each long line is accessed every sixth tile. Similarly, the vertical long line bus is designated LVN at the north end and LVS at the south end of FIG. 2. LVN and LVS extend through an offset region 421 similar to offset region 422. The long line buses extend the entire length or width of the tile array.
[0050] Another bus extends horizontally through the programmable region PROG-[0050] 2, a tristate buffer bus, 4 lines wide, designated TW at the west end and TE at the east end of FIG. 2. TW and TE are connected by programmable offset region 423. Programmable offset region 423 is similar to offset regions 421 and 422 in that it offsets the bus by one line; however, programmable offset region 423 is only four lines wide and also includes one programmable bidirectional PIP for extending the length of bus TW or TE. (The internal structure of programmable offset region 423 is shown in FIG. 3.) This PIP can be used either to “break” each line of the tristate bus every fourth tile, or to connect lines together every fourth tile to continue a tristate line up to the full width of the tile array. Two examples of tristate buses that can be used are described by Steven P. Young et al in commonly assigned U.S. Pat. No. 5,677,638, entitled “High Speed Tristate Bus with Multiplexers for Selecting Bus Driver”, issued Oct. 14, 1997, and commonly assigned U.S. Pat. No. 5,847,580, “High Speed Bidirectional Bus with Multiplexers”, by Bapat et al, issued Dec. 8, 1998, both of which are incorporated herein by reference. Programmable Bus Connections
[0051] Programmable bus connections are located in region INTERCONNECT-[0051] 2, which is part of programmable region PROG-2. Where hex buses cross single-length buses, open triangles indicate programmable connections in which one of the hex lines at the base of the triangle can drive one of the single-length lines at the tip of the triangle. Triangles pointing to hex lines represent buffered connections, and triangles pointing to single-length lines represent unbuffered connections such as simple pass transistors.
[0052] Note that not all intersections have triangles. For example, hex bus segment [0052] 6V can connect to east and west single-length buses E and W, but not to north and south buses N and S. Similarly, hex bus segment 6H can connect to north and south single-length buses N and S, but not to east and west buses E and W. Hex bus segment 6S coming from the south can connect to east and west single-length buses E and W and also to north bus N. This partial connectivity saves silicon area while still providing useful connections for good routability and performance. Similarly, hex bus segment 6N coming from the north can connect to east and west single-length buses E and W and also to south bus S; hex bus segment 6W coming from the west can connect to north and south single-length buses N and S and also to east bus E; and hex bus segment 6E coming from the east can connect to north and south single-length buses N and S and also to west bus W. In each of these cases, the connection may be limited by hardware or software to being unidirectional such that the hex buses can drive the single-length buses, but the single-length buses cannot drive the hex buses. (In the actual hardware of this embodiment, the connection is a bidirectional pass-gate; whereby the single-length buses can drive the hex buses. However, doing so reduces the speed of the resulting signal path dramatically, and therefore software for programming the FPGA does not select such a path.)
[0053] All of the single-length buses can be driven by output multiplexer OMUX-[0053] 2 by way of output bus OUT. Hex bus segments 6E, 6W, 6N, and 6S can also be driven by output bus OUT.
[0054] Vertical hex bus segments [0054] 6N and 6S can be connected together via programmable bus connection 404. (The internal structure of programmable bus connection 404 is shown in FIG. 4.) In this fashion, a line in a vertical hex bus can be continued 12, 18 or more tiles, rather than terminating after 6 tiles. Similarly, horizontal hex bus segments 6W and 6E can be connected together via programmable bus connection 405. Additionally, vertical hex bus segments 6N and 6S can be connected to horizontal hex bus segments 6W and 6E via programmable bus connections 406, 407, 408, and 409. The connectivity of these bus connections is shown in FIG. 4. Note also that hex bus segment 6V (which implements the connection to the third tile in the vertical direction) can drive horizontal hex bus segments 6W and 6E. Similarly, hex bus segment 6H (which implements the connection to the third tile in the horizontal direction) can drive vertical hex bus segments 6N and 6S. By this means, the direction of a signal path is easily changed between horizontal and vertical.
[0055] Horizontal hex bus segments [0055] 6E and 6W in a given tile can be driven by the two horizontal long lines accessed in the same tile. Vertical hex bus segments 6N and 6S in a given tile can be driven by the two vertical long lines accessed in the same tile. This capability allows a signal on a long line to fan out to an entire row or column of tiles with high speed, as later shown in FIGS. 10-12, by routing from the long line to each hex line (and hence to single lines) rather than propagating serially along the hex lines.
[0056] Switching structure [0056] 403 allows each single-length bus (N, S, E, W) to be connected to each of the other single-length buses. The connectivity in switching structure 403 is shown in FIG. 4, and is also described and shown in detail in FIG. 6A of commonly assigned U.S. Pat. No. 5,828,230 entitled “FPGA Two Turn Routing Structure with Lane Changing and Minimum Diffusion Area” by Steven P. Young, which is hereby incorporated herein by reference. Significance of Programmable Bus Connections
[0057] The bus connections in FIG. 2 implement an interconnect structure for FPGAs according to the invention. The buses of this embodiment interact in the following ways: 1) long lines can drive hex buses parallel thereto, but cannot drive either perpendicular hex buses or single-length buses except through the parallel hex buses; 2) hex buses can drive single-length buses both parallel and perpendicular thereto, but single-length buses cannot drive hex buses (although in one embodiment hex buses and single-length buses are connected through bidirectional PIPs, so a software limitation is imposed to make these connections unidirectional); 3) hex buses can further drive other hex buses both parallel and perpendicular thereto; and 4) most hex buses cannot connect to the inputs of input multiplexer IMUX-[0057] 2, although a limited number of lines from the hex buses can make such connections, as is later described. Therefore, in the interconnect structure of FIG. 2, it is preferred to make connections between tiles by placing a signal first on the longest suitable line and from there moving the signal to a shorter line. For example: 1) a connection to a logic block three or six tiles away would be made through a hex line, switching to a single-length line to reach the input multiplexer; 2) a connection to a logic block seven tiles away would be made through first a hex line and then a single-length line (this connection requires no more PIPs than the first example); 3) a connection to a logic block on the other side of the FPGA would be made through first a long line, then a hex line, then transferred to one or more single-length lines before entering the destination input multiplexer. In this manner, large RC trees are avoided, by ensuring that unbuffered PIPs do not drive heavily loaded lines. Further, the various types of signal paths throughout the interconnect structure have similar delays, increasing the overall performance of the circuit.
[0058] Of course, many signal paths do not travel only within one column or one row of tiles. To turn a corner, a signal path can move from a given single-length line to a perpendicular single-length line through the switching structure [0058] 403, or from a given hex line to a perpendicular hex line. Programmable Logic Portion of the Tile
[0059] The interconnect portion INTERCONNECT-[0059] 2 of the repeatable tile of FIG. 2 exists only for the purpose of delivering signals between the logic blocks in different tiles. The logic block LB of this embodiment comprises a configurable logic element CLE-2, an output multiplexer OMUX-2, an input multiplexer IMUX-2, and two tristate buffers 445. (The number of tristate buffers can of course vary in other embodiments of the invention.) Configurable logic element CLE-2, input multiplexer IMUX-2, and output multiplexer ONUX-2 in FIG. 2 have small solid black triangles on each input and output. The direction of the apex of the triangle indicates the direction of the signal flow. Thus, a triangle with its apex inside the rectangular box indicates an input line or bus and a triangle with its apex outside the rectangular box indicates an output line or bus.
[0060] Input multiplexer IMUX-[0060] 2 of this embodiment accepts inputs from: 1) each of single-length buses N, S, E, W; 2) four lines 428 of each 12-line vertical hex bus 42, 43, 44, 45, 46, 47; 3) four lines 429 from 8-line output bus OUT from output multiplexer OMUX-2; and 4) a total of four fast-feedback lines 430 from configurable logic element CLE-2. Input multiplexer IMUX-2 can programmably create the following output signals: 1) two 13-line buses 436 that provide input signals to configurable logic element CLE-2; 2) two signal lines 431 that drive the data inputs of tristate buffers 445; 3) two signal lines 432 that enable tristate buffers 445; 4) two lines of vertical long line bus LVS; and 5) two lines of horizontal long line bus LHE.
[0061] Configurable logic element CLE-[0061] 2 accepts only inputs from input multiplexer IMUX-2, and drives 12 output lines shown in FIG. 2 as two 6-line buses 433. Four of these output lines form fast-feedback paths 430 to input multiplexer IMUX-2. Outputs 433 from configurable logic element CLE-2 drive output multiplexer OMUX-2. OMUX-2 gets one additional input 434 from tristate buffer bus TW, and drives 8-line output bus OUT. Each of the two tristate buffers 445 drives two of the tristate lines in tristate buffer bus TW via 2-line bus 435. The connection of tristate buffers 445 to tristate buffer bus TW is shown in more detail in FIG. 3. Detailed Repeatable Tile Representation
[0062] FIGS. [0062] 3-12 show the preferred embodiment of a repeatable interconnect structure according to the invention, which embodiment is similar to the embodiment of FIG. 2. FIG. 3 shows a detailed representation of a single tile in the preferred embodiment. All features shown in FIG. 2 are represented, and some additional features are included. Labels in FIG. 3 correspond to the same labels in FIG. 2; thus corresponding structures are easily identified. There is one difference in representation: in FIG. 3, 72-line vertical routing track 6VN/6VS of FIG. 2 is represented as two 36-line vertical routing tracks, 6RN/6RS on the right and 6LN/6LS on the left. Half of the lines of each 12-line hex bus are drawn on each side of programmable region PROG, which corresponds to programmable region PROG-2 in FIG. 2. This representational difference has no effect on the functionality of the tile. In FIG. 3, the horizontal and vertical hex buses are shown with first lines of all buses are grouped together. Only the first lines are labeled. For example, labels 41 through 47 represent the first lines of hex bus segments 41 through 47 of FIG. 2. Hex bus segments 31 through 37 of FIG. 2 are similarly represented, but are not labeled in FIG. 3.
[0063] Interconnect region INTERCONNECT of FIG. 3 is functionally the same as interconnect region INTERCONNECT-[0063] 2 of FIG. 2. Features in the embodiment of FIG. 3 that are not shown in FIG. 2 include: 1) horizontal four-line global bus G along the top of the tile provides four global inputs to input multiplexer IMUX, which global inputs can be used to distribute high-fanout or high-speed global clock signals; 2) two direct output connection lines DOW from output bus OUT (driven by output multiplexer OMUX) exit the tile of FIG. 3 on the west edge, where they are connected to input lines DIE of an adjacent tile to the west; 3) two direct input connection lines DIE from the adjacent tile to the east enter the tile of FIG. 3 on the east edge and go to input multiplexer IMUX; 4) two direct output connection lines DOE from output bus OUT exit the tile of FIG. 3 on the east edge, where they are connected to lines DIW of an adjacent tile to the east; 5) two direct input connection lines DIW from the adjacent tile to the west enter the tile of FIG. 3 on the west edge and go to input multiplexer IMUX; and 6) two carry chains are included in configurable logic element CLE, each carry chain having an input on the south edge of the tile and an output on the north edge of the tile, the inputs designated CIN0 and CIN1 and the outputs designated COUT0 and COUT1, respectively. Some embodiments of the interconnect structure according to the invention do not have these features. Interconnect Region
[0064] FIG. 4 shows interconnect region INTERCONNECT of FIG. 3 in detail. All input and output lines are located in the same relative positions in FIGS. 3 and 4. Input and output lines to interconnect region INTERCONNECT in FIG. 3 can be identified by looking at the corresponding labels in FIG. 4. With the exception of the division of the vertical routing track, as described above, FIG. 4 also corresponds directly to the interconnect region INTERCONNECT-[0064] 2 of FIG. 2. Numerical labels identify features of the implementation previously shown and described. The connectivity in switching structure 403 is also described and shown in detail in FIG. 6A of commonly assigned U.S. Pat. No. 5,828,230 entitled “FPGA Two Turn Routing Structure with Lane Changing and Minimum Diffusion Area” by Steven P. Young, which is referenced above.
[0065] Programmable bus connections [0065] 404 and 405 in FIG. 4 demonstrate an additional feature of the invention. In programmable bus connections 404 and 405, vertical and horizontal hex buses (respectively) are programmably connected together to enable the continuation of a signal path in the same direction from one hex bus to the next. As shown in FIG. 4, some of these connections can be bidirectional (for example, implemented as pass transistors) while others are unidirectional (such as buffered connections). In one embodiment, the bidirectional hex buses (hex buses that can be programmably driven from either end) have more programmable connections to perpendicular buses than do the unidirectional hex buses (hex buses that can be programmably driven from only one end). Such bidirectional connections will be most flexible, while such unidirectional connections will be faster, since the PIP can be implemented as a single buffer without a programmable enable.
[0066] Vertical long lines LV[0066] 0 and LV6 identify the first and seventh lines, respectively, of vertical long line bus LVS that are accessed in each repeatable tile, as previously described. Similarly, horizontal long lines LH0 and LH6 identify the first and seventh lines, respectively, of the horizontal long line bus LHE.
[0067] As previously described, hex bus segments [0067] 6N, 6V, and 6S of FIG. 2 are divided in FIG. 3 into hex bus segments 6NR, 6VR, and 6SR, which exit the INTERCONNECT region on the east edge to form 36-line vertical routing track 6RN/6RS of FIG. 3, and hex bus segments 6NL, 6VL, and 6SL, which exit the INTERCONNECT region on the west edge to form 36-line vertical routing track 6LN/6LS of FIG. 3. Output Multiplexer OMUX
[0068] FIGS. 5, 6 and [0068] 7 show the output multiplexer (OMUX), configurable logic element (CLE) and input multiplexer (IMUX), respectively, of the embodiment shown in FIG. 3. All input and output lines are physically located in the same relative positions in FIGS. 3, 5, 6 and 7, thus input and output lines may readily be traced between the figures. The PIP patterns shown in FIGS. 5 and 7 form only one of many possible patterns. The implementation of these three blocks, and the connections between the blocks, may differ widely in different embodiments of the invention, provided that the three blocks are compatible with each other. For example, the number of connections between the blocks may be increased or decreased. Further, two or more of these blocks may be combined, or one or both of the input or output multiplexers may be eliminated. Although these variations are not further described herein, such modifications are encompassed by the inventive concept. It will be apparent to one skilled in the art after reading this specification that the present invention may be practiced within these and other architectural variations.
[0069] FIG. 5 shows details of the interior of output multiplexer OMUX. As shown in FIG. 5, the input lines (the outputs of logic block CLE) are selectively programmed to drive a series of eight outputs (OUT[0069] 0 to OUT7). Not all inputs can drive all outputs; a selection of how many PIPs to provide and where they should be located is typically made based on the trade-off constraints of performance, silicon area, and routing flexibility. Configurable Logic Element CLE
[0070] FIG. 6 shows the input and output signals for the configurable logic element (CLE) of the embodiment shown in FIG. 3. Many configurable logic elements can be used with the interconnect structure of the invention. For example, pages 4-11 through 4-23 of the Xilinx 1996 Data Book entitled “The Programmable Logic Data Book”, available from Xilinx, Inc., 2100 Logic Drive, San Jose, Calif. 95124, also incorporated herein by reference, describe a configurable logic element used in the XC4000 Series FPGA architecture which is compatible with the invention. Pages 4-294, 4-295, and 13-13 through 13-15 of the same document, also incorporated herein by reference, describe another compatible configurable logic element, used in the XC3000 Series FPGA architecture. Pages 4-188 through 4-190 of the same document, also incorporated herein by reference, describe yet another compatible configurable logic element, used in the XC5200 Family FPGA architecture. [0070]
[0071] One CLE that can be used with the embodiment of FIG. 3 is shown in FIGS. 6A and 6B. In this embodiment, the CLE is implemented in two similar portions called “slices”. To view the entire CLE, FIGS. 6A and 6B, each showing one slice, must be viewed together as shown in the key at the top of FIG. 6A. Programmable (configurable) bits controlling the function of the CLE are shown as small boxes containing an “x”. [0071]
[0072] The details of one function generator compatible with the interconnect structure of the invention are shown and described in commonly assigned, co-pending U.S. application Ser. No. 08/786,818 entitled “Configurable Logic Block with AND Gate for Efficient Multiplication in FPGAs” by Chapman et al, which is referenced above and incorporated herein by reference. [0072]
[0073] One CLE compatible with the interconnect structure of the invention is the CLE of FIGS. 6A and 6B, which includes four function generators: F and G in FIG. 6B, and H and J in FIG. 6A. Function generator F has four inputs F[0073] 1, F2, F3, F4; function generator G has four inputs G1, G2, G3, G4; function generator H has four inputs H1, H2, H3, H4; and function generator J has four inputs J1, J2, J3, J4. Each of function generators F, G, H, J also has a data-in input DI and an output O. In addition, each function generator F, G has a write strobe WS input WSF, WSG, respectively, generated in write strobe generator WB from clock input signal CK0 optionally inverted by multiplexer 69B, set/reset input signal SR0 optionally inverted by multiplexer 60B, and input signal BF optionally inverted by multiplexer VF. Each function generator H, J has a write strobe WS input WSH, WSJ, respectively, generated in write strobe generator WA from clock input signal CK1 optionally inverted by multiplexer 69A, set/reset input signal SR1 optionally inverted by multiplexer 60B, and input signal BH optionally inverted by multiplexer VH.
[0074] Function generators F, G, H, J of FIGS. 6A and 6B preferably include novel features described by Bauer in commonly assigned U.S. Pat. No. 5,889,413, entitled “Lookup Tables Which Double as Shift Registers”, issued Mar. 30, 1999, which is incorporated herein by reference. Briefly, these function generators have the ability to operate as shift registers as well as small RAMs and lookup tables. Write strobe signal WS controls the shifting function and the RAM write cycle. (In one embodiment, write strobe signal WS is implemented as two separate control signals, as described by Bauer.) Data-in port DI carries a data signal (in this embodiment derived from input signals BF, BH, BG, BJ) for writing to an addressed memory cell. When function generators F, G, H, J are configured as shift registers as described by Bauer, the shift register data input signal is taken from BF, BG, BH, BJ, respectively. Write strobe generators WB, WA generate write strobe signals WSF and WSG, WSH and WSJ, respectively. Signals WSF, WSG, WSH, WSJ strobe data into their respective function generators during a write cycle when the function generators are configured as RAM, and cause a shift when the function generators are configured as shift registers. The use of function generators as RAM is described in pages 4-11 through 4-23 of the Xilinx 1996 Data Book entitled “The Programmable Logic Data Book”, referenced above and incorporated herein by reference. [0074]
[0075] The function generator of this embodiment can therefore be configured as a look-up table, a shift register, a 16×1 RAM, half of a 16×1 dual-ported RAM (when the two function generators in a slice are paired together), half of a 16×2 RAM, or half of a 32×1 RAM (when the two function generators in a slice are paired together). Not all combinations of function generator configurations are supported in a single slice. The supported combinations are detailed below. [0075]
[0076] In this embodiment, write strobe signals WSF and WSG are controlled by write strobe generator WB. A High (logic 1) pulse on function generator input WS causes a write to occur to the function generator look-up table, to either the first memory cell of the look-up table (if configured as a shift register) or to an addressed cell (if configured as a RAM). (See commonly assigned U.S. Pat. No. 5,889,413 entitled “Lookup Tables Which Double as Shift Registers”, which is referenced above.) [0076]
[0077] Write strobe generator WB functions as follows in the supported combinations of function generator configurations: a) when function generators F and G are both configured as look-up tables, write strobe signals WSF and WSG are held Low (logic 0); b) when function generator G is configured as a 16×1 RAM or shift register and function generator F is configured as a look-up table, write strobe signal WSF remains Low while write strobe signal WSG pulses High when SR[0077] 0 is active (Low or High, depending on whether multiplexer 60B is inverting or not) and CK0 makes the proper transition (falling or rising, depending on whether multiplexer 69B is inverting CK0 or not); c) when function generators F and G are configured as a dual-ported 16×1 RAM, a 16×2 RAM, or a pair of shift registers, write strobe signals WSF and WSG pulse High together when SR0 is active and CK0 makes the proper transition; d) when function generators F and G are configured as a single 32×1 RAM, one but not both of write strobe signals WSF and WSG pulses High, when SR0 is active and CK0 makes the proper transition. The generated High pulse occurs on write strobe signal WSF if multiplexer VF is non-inverting and input signal BF is High or if multiplexer VF is inverting and input signal BF is low, otherwise the High pulse occurs on write strobe signal WSG. AS can be seen from the above description, SR0 acts as a write enable and CK0 acts as a clock to a synchronous RAM or shift register, while BF serves as a fifth address bit when the two function generators in a slice are configured as a 32-bit RAM.
[0078] Write strobe signals WSH and WSJ are generated by write strobe generator WA and controlled by input signals SR[0078] 1, CK1, and BH, in a corresponding fashion.
[0079] AND gates [0079] 61F, 61G, 61H, 61J and multiplexers 81F, 81G, 81H, 81J, CF, CG, CH, CJ implement multiplier and carry logic functions as described in commonly assigned, co-pending U.S. patent application Ser. No. 08/786,818 entitled “Configurable Logic Block with AND Gate for Efficient Multiplication in FPGAs” by Chapman et al, which is referenced above and incorporated herein by reference. In FIGS. 6A and 6B, each of multiplexers 81F, 81G, 81H, 81J has separate programmable memory cells that determine the multiplexer output. In another embodiment, the two multiplexers in each slice (81F and 81G, 81H and 81J) share memory cells and therefore have the same programmed functionality.
[0080] In the embodiment of FIGS. 6A and 6B, each slice has a separate carry chain. One carry chain is from input signal CIN[0080] 0 to output signal COUT0 and incorporates the outputs of function generators F and G. The other carry chain is from input signal CIN1 to output signal COUT1 and incorporates the outputs of function generators H and J. In one embodiment, the two carry chains are implemented as described in U.S. Pat. No. 5,629,886, entitled “Method and Structure for Providing Fast Propagation of a Carry Signal in a Field Programmable Gate Array” by Bernard J. New, issued May 13, 1997, which is referenced above and incorporated herein by reference.
[0081] Each slice further includes five-input-function multiplexer F[0081] 5A, F5B and six-input-function multiplexer F6A, F6B. In FIG. 6A, multiplexer F6A receives an input signal on west-going line F5W from the output of multiplexer F5B in FIG. 6B. In FIG. 6B, six-input-function multiplexer F6B receives an input signal on east-going line F5E from the output of multiplexer F5A in FIG. 6A. Multiplexer F6A receives a second input signal from multiplexer F5A, and multiplexer F6B receives a second input signal from multiplexer F5B. Multiplexer F5A receives inputs from outputs O of function generators H and J. Multiplexer F5B receives inputs from outputs O of function generators F and G.
[0082] The CLE of this embodiment receives four extra control signals on lines BF, BH, BG, and BJ. These lines can be programmably inverted using multiplexers VF, VH, VG, and VJ, respectively, and control several useful functions. The five-input function multiplexers F[0082] 5A, F5B are controlled by lines BH, BF, respectively. The six-input function multiplexers F6A, F6B are controlled by lines BJ, BG, respectively. A user can form a four-input multiplexer using multiplexer F5A and the outputs of function generators H and J, with function generators H and J each configured as a two-input multiplexer. Similarly, a four-input multiplexer can be formed using multiplexer F5B and the outputs of function generators F and G. A user can form an eight-input multiplexer using either multiplexer F6A or multiplexer F6B receiving inputs from both multiplexers F5A and F5B, each of which is receiving outputs from its respective function generators. Further, with multiplexer F5A, the memory cells in function generators H and J can operate as a 32-bit RAM. With multiplexer F5B, the memory cells in function generators F and G can operate as a 32-bit RAM. Lines BF, BH, BG, BJ serve as multiplexer control lines or RAM address lines, in addition to other functions as later described.
[0083] Registers RX, RY, RZ, RV generate outputs XQ, YQ, ZQ, VQ, respectively. Clock inputs to registers RX, RY, RZ, RV are designated by open triangles with apexes pointing into these registers. Clock enable inputs CE can be disabled using AND-gates [0083] 62A, 62B. (AND-gates 62A, 62B are controlled by a programmable memory cell that provides one of the inputs to each AND-gate.) Registers RX, RY, RZ, RV can be set or reset either synchronously or asynchronously through set port S or reset port R. (The choice of synchronous or asynchronous is made by a programmable memory cell shared between the two registers in each slice.) Set/reset input signals SR0, SR1 are controlled by multiplexers 60A, 60B and AND-gates 67A, 67B. Inputs SR0, SR1 (if enabled by AND-gates 67A, 67B) can set or reset the registers. Set/reset control units R81B, R82B, R81A, R82A are programmably controlled to either set or reset registers RX, RY, RZ, RV, respectively. Registers RX, RY, RZ, RV can also be configured as latches, the choice being made by a programmable memory cell shared between the two registers in each slice.
[0084] XOR gates SF, SG, SH, SJ generate the sum portion of an addition or multiplication, as described in commonly assigned U.S. Pat. No. 5,889,413 entitled “Lookup Tables Which Double as Shift Registers”, which is referenced above. [0084]
[0085] Multiplexers MF, MG, MH, MJ select between signals to place onto output lines X, Y, Z, V, respectively. Multiplexers MF and MH select between the related function generator output signal, sum output signal, and five-input function multiplexer output signal. Multiplexers MG and MJ select between the related function generator output signal, sum output signal, and six-input function multiplexer output signal. The function generator F, G, H, J output signal is selected when a logical function of up to four inputs is chosen. The five-input function multiplexer F[0085] 5A, F5B output signal is selected when one of a limited set of nine-input functions is chosen, including any logical function of five inputs, a 4-to-1 multiplexer, and any other function that can be implemented as two four-input functions feeding a 2-to-1 multiplexer. The six-input function multiplexer F6A, F6B output signal is selected when one of a limited set of 18-input functions is chosen, including any logical function of six inputs, an 8-to-1 multiplexer, and any other function that can be implemented as the output of two five-input function multiplexers F5A, F5B feeding a 2-to-1 multiplexer. The sum output signal of XOR gates SF, SG, SH, SJ is selected when an appropriate arithmetic operation such as addition, subtraction or multiplication is implemented.
[0086] Multiplexers OF, OG, OH, OJ allow the carry chain to bypass the related function generator. Multiplexers OF, OG, OH, OJ select whether the respective carry multiplexer CF, CG, CH, CJ is controlled by the function generator output signal 0 or whether the carry chain bypasses the function generator by propagating the carry-in signal to the next stage along the carry chain. Multiplexers CA, CB allow additional access onto their respective carry chains by selecting either a carry-in signal CIN[0086] 0, CIN1 (buffered in this embodiment by inverters I121B and I122B, I121A and I122A, respectively) or an input signal BF, BH as the starting point for the carry chain in the CLE.
[0087] Multiplexers DF, DH select a signal to be loaded into related registers RX, RZ either from the outputs of multiplexers MF, MH, or from input signals BF, BH, respectively. Multiplexers DG, DJ select a signal to be loaded into related registers RY, RV either from the outputs of multiplexers MG, MJ, or from input signal lines BG, BJ, respectively. [0087]
[0088] Multiplexers BB, BA are bypass multiplexers that can apply carry-out signals COUT[0088] 0, COUT1 to output lines YB, VB, or can forward input signals BG, BJ to output lines YB, VB to route a signal generated on one side of the CLE to a destination on the other side of the CLE.
[0089] AND-gates BRB, BRA allow input signals BG, BJ to serve as set/reset signals to registers RX, RY, RZ, RV, or alternatively allows input signals BG, BJ to be decoupled from registers RX, RY, RZ, RV so that input signals BG, BJ can be used for other purposes, such as controlling multiplexers F[0089] 6B, F6A or providing a DI signal to function generators G, J, without setting or resetting the registers. A logic 0 in the memory cells controlling AND-gates BRB, BRA decouples lines BG, BJ from set/reset control units R81B, R82B, R81A, R82A; while a logic 1 in the associated memory cell enables AND-gates BRB, BRA to use the signals on BG, BJ to set or reset registers RX, RY, RZ, RV as determined by set/reset control units R81B, R82B, R81A, R82A. Input Multiplexer IMUX
[0090] FIG. 7 shows details of the interior of input multiplexer IMUX of the embodiment shown in FIG. 3. As shown in FIG. 7, the input lines in this embodiment (G, [0090] 430, DIW, 428, DIE, 429) are selectively programmed to drive a series of 34 outputs (431, 432, 436, LV0, LV6, LH0, LH6). 26 of these outputs drive CLE inputs 436, two drive the data inputs 431 and two drive the enable inputs 432 of tristate buffers 445, two drive LV0 and LV6 (two lines of vertical long line bus LVS), and two drive LH0 and LH6 (two lines of horizontal long line bus LHE). Not all inputs can drive all outputs; a selection of how many PIPs to provide and where they should be located is typically made based on the trade-off constraints of performance, silicon area, and routing flexibility. In one embodiment, any IMUX output not connected to any IMUX input line is connected through a pullup to the positive voltage supply. Routing Examples
[0091] FIGS. 8 and 9 show examples of signal paths implemented in an array of tiles according to the embodiment shown in FIG. 3. The labeling convention used in FIGS. 8 and 9 is that tiles are identified by a row and column number based on the portion of the array shown in the figure. Within a tile the line labels correspond to the bus labels indicated in FIG. 2. However, in FIGS. 8 and 9, the label indicates a single line rather than the entire bus. Only lines used in the examples are shown and labeled. With this labeling convention, some labels are used more than once. But no ambiguity exists, because all tiles are identical and each tile has a different tile number. The tile numbers distinguish between labels used more than once in the same drawing. [0091]
[0092] FIG. 8 shows a routing example of a signal path using hex and single-length lines. A signal is sourced by source logic block L[0092] 0 of tile 8-2, and programmably connected within the tile through line OUT to hex segment 6N. Hex segment 6N turns north and becomes vertical hex segment 47. This signal must fan out to logic blocks at three destinations in tiles 5-3, 4-3, and 1-2. The signal passes through tiles 7-2 through 3-2 on segments 46 through 42. At tile 5-2, the signal is on segment 44, where it fans out to segment 6V. In tile 5-2, segment 6V is programmably connected to single-length line E, which continues east into tile 5-3 as line W. In tile 5-3, line W is an input to the first destination, logic block L1, and also to switching structure 403, where it is programmably connected to line N. Line N of tile 5-3 continues north into tile 4-3 as line S, which is an input to the second destination, logic block L2. The northern end of the vertical hex line is reached in tile 2-2, where the signal is on segment 41. Segment 41 is programmably connected to single-length line N, which continues north into tile 1-2 as line S. In tile 1-2, line S is an input to the third destination, logic block L3.
[0093] Thus in FIG. 8 a signal net is formed from source logic block L[0093] 0 to destination logic blocks L1, L2 and L3. The connections are fast because the PIP onto the hex line is buffered, there is a low capacitance on the hex line due to the small number of PIPs on the hex line, and each path from source to destination has a small total number of PIPs inserting delay into the signal path. In the example of FIG. 8, each signal path between source and destination logic blocks requires at most three PIPs: one buffered PIP onto the hex line and one or two unbuffered PIPs onto single-length lines. (A logic block is the structure designated LB in FIG. 2.)
[0094] FIG. 9 shows a routing example of a signal path using a long line, hex lines, and single-length lines. A signal is sourced by source logic block L[0094] 8 of tile 1-1, and must be routed to logic blocks L4, L5, L6, and L7. The signal from logic block L8 is programmably connected within tile 1-1 to one of the vertical long lines LV accessible from that particular tile. Long line LV extends southward for 18 tiles to tile 19-1, where long line LV is programmably connected to hex segment 6S, then continues southward to drive other hex lines (not shown) south of the portion of the array shown in FIG. 9. (The gap of 18 tiles in FIG. 9 is included to demonstrate that long line LV must carry the signal far enough that a long line is more efficient than concatenated hex lines.) From hex segment 6S in tile 19-1, the signal turns south and becomes vertical hex segment 41. The signal passes through tiles 20-1 through 24-1 on segments 42 through 46. At tile 22-1, the signal is on segment 44, where it fans out to segment 6V and continues south to tile 25-1. From segment 6V in tile 22-1, the signal makes a programmable connection to segment 6E (demonstrating the connection of two perpendicular hex lines). The signal then moves east on segment 37, traversing tiles 22-2 through 22-6 on segments 36 through 32. At tile 22-4, the signal is on segment 34, where it fans out to segment 6H and continues east to tile 22-5. In tile 22-4, the signal makes a programmable connection between segment 6H and single-length line S, which continues south into tile 23-4 as line N. In tile 23-4, line N enters switching structure 403 and connects to line E, which continues east into tile 23-5 on line W, an input to the first destination, logic block L4. From segment 32 in tile 22-6, the signal continues east into tile 22-7 on segment 31, then turns north onto segment 6W. Segment 6W is programmably connected to line N, which continues north into tile 21-7 as line S and is an input to the second destination, logic block L5.
[0095] Returning to the vertical hex line, the signal on segment [0095] 46 in tile 24-1 continues south into tile 25-1 on segment 47, which turns west on segment 6N. From segment 6N in tile 25-1, the signal makes a programmable connection to segment 6E. The signal then moves east on segment 37, traversing tiles 25-2 through 25-6 on segments 36 through 32. From segment 32 in tile 25-6, the signal continues east into tile 25-7 on segment 31, then turns north onto segment 6W. Segment 6W is programmably connected to line S, which continues south into tile 26-7 as line N and is an input to the third destination, logic block L6. One further connection remains to be made. From the south end of the vertical hexline, in tile 25-1, the signal on segment 6V makes an additional programmable connection to line E. Line E continues east into tile 25-2 on line W. In tile 25-2, line w enters switching structure 403 and connects to line S, which continues south into tile 26-2 on line N, an input to the fourth destination, logic block L7.
[0096] Thus in FIG. 9 a signal net is formed from source logic block L[0096] 8 to destination logic blocks L4, L5, L6 and L7. As in the example of FIG. 8, the connections to long lines and hex lines are buffered and the RC delay of the unbuffered connection is minimized. In this example, these destinations required at most four PIPs in the interconnect path between logic blocks: one or two buffered PIPs onto hex lines, and one or two unbuffered PIPs onto single-length lines. Yet 31 tiles were traversed between source logic block L8 and destination logic block L6.
[0097] In all of the above routing examples, alternative paths are available, many of which require more PIPs than the signal paths used in these examples. If some of the interconnect lines on one path are already used by other signals, an alternative path may be used. [0097] High Fanout Routing Examples
[0098] FIG. 10 shows a routing example of a high fanout control signal, such as a clock (CK), clock enable (CE), set/reset (SR), or tristate enable signal ([0098] 432) delivered from a vertical long line to a column of tiles. A signal can be placed onto a vertical long line either from outside the tile array, or by using one of lines 437I of FIG. 2. From vertical long line bus LV, in every twelfth tile (TILE4) the high fanout control signal emerges on line 600. Line 600 drives two vertical hex lines, one of which (H1) extends north six tiles from TILE4 and one of which (H2) extends south six tiles from TILE4. From hex lines H1 and H2, lines 428 (shown in FIG. 2) permit four bits of each vertical hex bus to drive the logic block. The input multiplexer (IMUX) of FIG. 7 permits vertical hex lines, through lines 428, to drive control signals CK, CE, SR, and 432 without passing through a single-length line, although signals on other CLE inputs, such as function generator inputs, must always pass through a single-length line to reach the CLE. Hex line H1 drives the logic block in TILE4, plus the five tiles adjacent to the north. Note that in the northernmost tile reached by hex line H1, line 428 does not make contact with hex line Hl. Instead, contact is made with hex line H3, which originates 12 tiles north of TILE4. Hex line H2 drives the six logic blocks adjacent to the south of TILE4. North of hex line H1, hex line H3 drives the adjacent six logic blocks. South of hex line H2, hex line H4 drives the adjacent six logic blocks. Each hex line drives a high fanout control signal in six logic blocks. The tiles accessed by hex lines H1 and H2 (labeled 610 in FIG. 10) form a column of 12 tiles that can be vertically repeated to access an entire column of an array of repeating tiles.
[0099] FIG. 10 shows that for a high fanout control signal, there is only one PIP between a signal on a vertical long line and any destination logic block in the same column. (However, in the preferred embodiment, there is an additional PIP added to the path in input multiplexer IMUX.) Therefore, there is very little delay on this path. Additionally, this distribution method has the same number of PIPs and the same number of interconnect lines (one long line, one hex line) in the path to each logic block input. Therefore, there is very low skew on this path. Although the global lines (G in FIG. 3) would typically be used for up to four clocks, the distribution method of FIG. 10, by taking advantage of the hierarchical nature of the interconnect scheme according to the invention, produces an excellent additional clock distribution network with high performance and low skew. [0099]
[0100] FIG. 11 shows a routing example of a high fanout function generator input signal (or other non-control signal) delivered from a vertical long line to a column of tiles. A signal can be placed onto a vertical long line either from outside the tile array, or by using one of lines [0100] 437I of FIG. 2. From vertical long line bus LV, in every twelfth tile (TILE5-6) the high fanout function generator input signal emerges on line 601. Line 601 drives two vertical hex lines, one of which (H5) extends north six tiles from TILE4 and one of which (H6) extends south six tiles from TILE4. From hex lines H5 and H6, line segments 6S, 6N and 6V are used in conjunction with single-length lines N, S and switching structures 403 to drive the logic blocks, similar to the interconnections in FIGS. 8 and 9. In one embodiment, signals connecting to function generator inputs inside the CLE must always pass through a single-length line to reach the CLE. Hex line H5 drives, through single-length lines, the logic blocks in tiles TILE5-7, TILE5-8, and TILE5-3. In addition, hex line H5 drives, through single-length lines, the logic blocks in the three southernmost tiles of the 12-tile column to the north of column 611 (corresponding to tiles TILE5-10, TILE5-11, and TILE5-12 of column 611). Hex line H6 drives, through single-length lines, the logic blocks in tiles TILE5-4, TILE5-5, TILE5-6, and TILE5-9. In addition, hex line H6 drives, through single-length lines, the logic blocks in the two northernmost tiles of the 12-tile column to the south of column 611 (corresponding to tiles TILE5-1 and TILE5-2 of column 611). Tiles TILE5-1 and TILE5-2 are driven, through single-length lines, by hex line H7, which originates 12 tiles north of TILE5-6. Tiles TILE5-10, TILE5-11, and TILE5-12 are driven, through single-length lines, by hex line H8, which originates 12 tiles south of TILE5-6. Each hex line drives a high fanout function generator input signal in six logic blocks. Thus, by cooperation between adjacent 12-tile columns 611, a repeating pattern is formed that enables the distribution of non-control signals to an entire column of an array of repeating tiles.
[0101] Note that other connections of hex lines and single-length lines are possible. For example, the logic block in TILE[0101] 5-2 can be driven by hex line 6V through a switching structure and single-length line in TILE5-3 (not shown) and hence through an additional switching structure and single-length line in TILE5-2 to drive the logic block in TILE5-1. This or a similar arrangement can be used in the northernmost implementation of 12-tile column 611, to access the two northernmost tiles. Similar connections can be made in the southernmost implementation of 12-tile column 611 to access the three southernmost tiles.
[0102] FIG. 11 shows that for a high fanout function generator input signal, there are either two or three PIPs between a signal on a vertical long line and any destination logic block in the same column: one from the long line to a hex line, one from the hex line to a first single-length line, and sometimes one to a second single-length line through the switching structure. (However, in the preferred embodiment, there is an additional PIP added to the path in input multiplexer IMUX.) Only the last one or two PIPs are unbuffered. Therefore, there are no large unbuffered RC trees on this net, no signal path has more than two unbuffered PIPs in series, and the delay on each signal path is minimized. [0102]
[0103] FIG. 12 shows an example of how to combine the interconnect of FIGS. [0103] 10 and/or 11 to deliver a high-fanout signal to an array of tiles. A signal can be placed onto a horizontal long line either from outside the tile array, or by using one of lines 438 of FIG. 2. The signal enters the portion of the array shown in FIG. 12 on a line in horizontal long line bus LH. Long line bus LH traverses a row of 12 tiles 612, in which it makes contact with 12 logic blocks through hex lines and single-length lines. The horizontal single-length lines connected to some horizontal hex lines can drive vertical long lines through PIPs in the IMUX. From the logic blocks in the 12 tiles of FIG. 12, lines 437I extend from the input multiplexer (see FIG. 2) in the logic block to drive the vertical long lines LV. Using lines 437I in each of the 12 tiles 612, the signal on horizontal long line bus LH is transferred to 12 vertical long lines LV. Columns of 12 vertical tiles (613) can be implemented by either 12 tiles 610 of FIG. 10 (for control signals) or 12 tiles 611 of FIG. 11 (for function generator or other non-control inputs). Multiple 12-tile columns 613 can be placed to increase the number of accessible tiles. In this manner, high fanout signals can be distributed from an originating tile to any tile in the array.
[0104] As an alternative to the distribution scheme of FIG. 12, special tiles can be created that permit, for example, hex lines to drive long lines, thus bypassing the single-length lines used in FIG. 12. Such special tiles can be advantageously placed in rows along the top and bottom edges of the tile array, or at intervals within the array. [0104] Summary
[0105] The interconnect structure of the present invention offers uniformly fast signals throughout the FPGA, due to the reduced loading (limited number of PIPs) on each interconnect line, the lack of large RC trees on signal nets, and the relatively small number of PIPs between source and destination on each path. The nearly uniform delays on various interconnect paths greatly mitigate a common situation in FPGAs where a few signal paths, much slower than the average signal path, significantly reduce the overall performance of the circuit. Intermediate-length lines enable the efficient implementation of localized circuits and ease the burden on routing software. The limited number of PIPs, and the judicious arrangement thereof, imparts a high degree of routing flexibility per silicon area consumed. Further routing flexibility is provided by the ability of logic block inputs to contact interconnect lines located on any edge of the logic block. Thus it will be understood that the present invention provides a new and useful interconnect structure for FPGAs. [0105]
[0106] Those having skill in the relevant arts of the invention will now perceive various modifications and additions which may be made as a result of the disclosure herein of the preferred embodiment. Accordingly, all such modifications and additions are deemed to be within the scope of the invention, which is to be limited only by the appended claims and their equivalents. [0106]

权利要求:
Claims (23)
[1" id="US-20010007428-A1-CLM-00001] 1. An interconnect structure for a programmable logic device, the programmable logic device including a plurality of substantially identical tiles, each tile including a logic block, the interconnect structure in each tile comprising:
a plurality of single-length buses for coupling the logic block to any adjacent logic blocks;
a plurality of vertical intermediate-length buses for coupling the logic block to non-adjacent logic blocks in a vertical direction;
a plurality of horizontal intermediate-length buses for coupling the logic block to non-adjacent logic blocks in a horizontal direction; and
a switching structure for selectively coupling the vertical intermediate-length buses and the horizontal intermediate-length buses to the logic block via the single-length buses.
[2" id="US-20010007428-A1-CLM-00002] 2. The interconnect structure of
claim 1 , wherein the switching structure includes a plurality of programmable interconnect points for coupling each line in at least one intermediate-length bus to N single-length buses, wherein N is less than the total number of single-length buses.
[3" id="US-20010007428-A1-CLM-00003] 3. The interconnect structure of
claim 1 , wherein the switching structure includes a plurality of programmable interconnect points for coupling each line in each intermediate-length bus to N single-length buses, wherein N is less than the total number of single-length buses.
[4" id="US-20010007428-A1-CLM-00004] 4. The interconnect structure of
claim 1 , wherein each vertical intermediate-length bus traverses at least two logic blocks without coupling to the at least two logic blocks.
[5" id="US-20010007428-A1-CLM-00005] 5. The interconnect structure of
claim 1 , wherein each horizontal intermediate-length bus traverses at least two logic blocks without coupling to the at least two logic blocks.
[6" id="US-20010007428-A1-CLM-00006] 6. The interconnect structure of
claim 1 , wherein at least one vertical intermediate-length bus includes a line that is abutting another line in another vertical intermediate-length bus, and wherein said switching structure includes a programmable connection for connecting the abutting lines.
[7" id="US-20010007428-A1-CLM-00007] 7. The interconnect structure of
claim 1 , wherein at least one horizontal intermediate-length bus includes a line that is abutting another line in another intermediate-length bus, and wherein said switching structure includes a programmable connection for connecting the abutting lines.
[8" id="US-20010007428-A1-CLM-00008] 8. The interconnect structure of
claim 1 , wherein the logic block includes an input multiplexer for receiving inputs from the single-length buses.
[9" id="US-20010007428-A1-CLM-00009] 9. The interconnect structure of
claim 8 , wherein the input multiplexer further receives inputs from one of the horizontal intermediate-length buses and the vertical intermediate-length buses.
[10" id="US-20010007428-A1-CLM-00010] 10. The interconnect structure of
claim 8 , further including a plurality of long lines, wherein the input multiplexer provides outputs to at least one of the plurality of long lines.
[11" id="US-20010007428-A1-CLM-00011] 11. The interconnect structure of
claim 1 , wherein the logic block includes an output multiplexer for providing outputs to the single-length buses.
[12" id="US-20010007428-A1-CLM-00012] 12. The interconnect structure of
claim 11 , wherein the switching structure includes a plurality of programmable connections for coupling the outputs to each single-length bus.
[13" id="US-20010007428-A1-CLM-00013] 13. A method of designing an interconnect structure for a programmable logic device, the programmable logic device including a plurality of substantially identical tiles, each tile including a logic block, the method comprising:
providing a plurality of single-length buses for coupling the logic block to any adjacent logic blocks;
providing a plurality of vertical intermediate-length buses for coupling the logic block to non-adjacent logic blocks in a vertical direction;
providing a plurality of horizontal intermediate-length buses for coupling the logic block to non-adjacent logic blocks in a horizontal direction; and
providing a switching structure for selectively coupling the vertical intermediate-length buses and the horizontal intermediate-length buses to the first logic block via the single-length buses.
[14" id="US-20010007428-A1-CLM-00014] 14. The method of
claim 13 , further including providing a plurality of programmable interconnect points in the switching structure for coupling each line in at least one intermediate-length bus to N single-length buses, wherein N is less than the total number of single-length buses.
[15" id="US-20010007428-A1-CLM-00015] 15. The method of
claim 13 , further including providing a plurality of programmable interconnect points in the switching structure for coupling each line in each intermediate-length bus to N single-length buses, wherein N is less than the total number of single-length buses.
[16" id="US-20010007428-A1-CLM-00016] 16. The method of
claim 13 , further including providing that each vertical intermediate-length bus traverses at least two logic blocks without coupling to the at least two logic blocks.
[17" id="US-20010007428-A1-CLM-00017] 17. The method of
claim 13 , further including providing that each horizontal intermediate-length bus traverses at least two logic blocks without coupling to the at least two logic blocks.
[18" id="US-20010007428-A1-CLM-00018] 18. The method of
claim 13 , further including providing a line in at least one vertical intermediate-length bus that is abutting another line in another vertical intermediate-length bus, and programmably connecting the abutting lines.
[19" id="US-20010007428-A1-CLM-00019] 19. The method of
claim 13 , further including providing a line in at least one horizontal intermediate-length bus that is abutting another line in another intermediate-length bus, and programmably connecting the abutting lines.
[20" id="US-20010007428-A1-CLM-00020] 20. The method of
claim 13 , further including providing means for receiving inputs from the single-length buses and selectively providing the inputs to the logic block.
[21" id="US-20010007428-A1-CLM-00021] 21. The method of
claim 13 , further including providing means for receiving inputs from one of the horizontal intermediate-length buses and the vertical intermediate-length buses and selectively providing the inputs to the logic block.
[22" id="US-20010007428-A1-CLM-00022] 22. The method of
claim 13 , further including providing a plurality of long lines and allowing the logic block to drive outputs to at least one of the plurality of long lines.
[23" id="US-20010007428-A1-CLM-00023] 23. The method of
claim 13 , further including driving a plurality of outputs from the logic block onto each single-length bus.

类似技术:

---

公开号 | 公开日 | 专利标题

---

US6448808B2|2002-09-10|Interconnect structure for a programmable logic device

---

US6204689B1|2001-03-20|Input/output interconnect circuit for FPGAs

---

US5907248A|1999-05-25|FPGA interconnect structure with high-speed high fanout capability

---

US6396303B1|2002-05-28|Expandable interconnect structure for FPGAS

---

US5942913A|1999-08-24|FPGA repeatable interconnect structure with bidirectional and unidirectional interconnect lines

---

US6069490A|2000-05-30|Routing architecture using a direct connect routing mesh

---

US6124731A|2000-09-26|Configurable logic element with ability to evaluate wide logic functions

---

US6084429A|2000-07-04|PLD having a window pane architecture with segmented and staggered interconnect wiring between logic block arrays

---

US5598109A|1997-01-28|Programmable logic array device with grouped logic regions and three types of conductors

---

US6239613B1|2001-05-29|Tristate structures for programmable logic devices

---

US5644496A|1997-07-01|Programmable logic device with internal time-constant multiplexing of signals from external interconnect buses

---

US5986465A|1999-11-16|Programmable logic integrated circuit architecture incorporating a global shareable expander

---

US5701091A|1997-12-23|Routing resources for hierarchical FPGA

---

US5894565A|1999-04-13|Field programmable gate array with distributed RAM and increased cell utilization

---

US5581199A|1996-12-03|Interconnect architecture for field programmable gate array using variable length conductors

---

US6130551A|2000-10-10|Synthesis-friendly FPGA architecture with variable length and variable timing interconnect

---

US5883526A|1999-03-16|Hierarchical interconnect for programmable logic devices

---

US6294928B1|2001-09-25|Programmable logic device with highly routable interconnect

---

US5621650A|1997-04-15|Programmable logic device with internal time-constant multiplexing of signals from external interconnect buses

---

WO1998038740A1|1998-09-03|Fpga having logic element carry chains capable of generating wide xor functions

---

US6882177B1|2005-04-19|Tristate structures for programmable logic devices

---

WO1999048004A1|1999-09-23|Sram bus architecture and interconnect to an fpga

---

US6150841A|2000-11-21|Enhanced macrocell module for high density CPLD architectures

---

GB2318663A|1998-04-29|Hierarchical interconnect for programmable logic devices

---

US6630842B1|2003-10-07|Routing architecture for a programmable logic device

同族专利:

---

公开号 | 公开日

---

US6292022B2|2001-09-18|

---

US20020008541A1|2002-01-24|

---

EP0972344A1|2000-01-19|

---

DE69716771D1|2002-12-05|

---

JP2001519133A|2001-10-16|

---

US6448808B2|2002-09-10|

---

EP0972344B1|2002-10-30|

---

DE69716771T2|2003-06-12|

---

WO1998043354A1|1998-10-01|

---

US6107827A|2000-08-22|

---

US5963050A|1999-10-05|

---

US6204690B1|2001-03-20|

---

JP3948497B2|2007-07-25|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

US20040261048A1|2003-06-19|2004-12-23|International Business Machines Corporation|Wire trimmed programmable logic array|

---

US20050218928A1|2004-03-30|2005-10-06|Pani Peter M|Scalable non-blocking switching network for programmable logic|

---

US20060023704A1|2004-07-29|2006-02-02|Pani Peter M|Interconnection fabric using switching networks in hierarchy|

---

US7242216B1|2004-11-08|2007-07-10|Herman Schmit|Embedding memory between tile arrangement of a configurable IC|

---

US7259587B1|2004-11-08|2007-08-21|Tabula, Inc.|Configurable IC's with configurable logic resources that have asymetric inputs and/or outputs|

---

US7276933B1|2004-11-08|2007-10-02|Tabula, Inc.|Reconfigurable IC that has sections running at different looperness|

---

US7282950B1|2004-11-08|2007-10-16|Tabula, Inc.|Configurable IC's with logic resources with offset connections|

---

US20070241777A1|2004-02-14|2007-10-18|Herman Schmit|Configurable Circuits, IC's and Systems|

---

US20070241772A1|2005-03-15|2007-10-18|Herman Schmit|Embedding memory within tile arrangement of a configurable ic|

---

US20070241781A1|2005-03-15|2007-10-18|Brad Hutchings|Variable width management for a memory of a configurable IC|

---

US20070241775A1|2004-11-08|2007-10-18|Jason Redgrave|Storage elements for a configurable ic and method and apparatus for accessing data stored in the storage elements|

---

US20070244959A1|2005-03-15|2007-10-18|Steven Teig|Configurable IC's with dual carry chains|

---

US20070244961A1|2004-11-08|2007-10-18|Herman Schmit|Configurable IC with configurable routing resources that have asymmetric Input and/or outputs|

---

US20070244958A1|2004-11-08|2007-10-18|Jason Redgrave|Configurable IC's with carry bypass circuitry|

---

US20070241778A1|2004-11-08|2007-10-18|Herman Schmit|IC with configurable storage circuits|

---

US20070241780A1|2004-11-08|2007-10-18|Steven Teig|Reconfigurable ic that has sections running at different reconfiguration rates|

---

US20070241787A1|2004-06-30|2007-10-18|Herman Schmit|Configurable Circuits, IC's, and Systems|

---

US20070241773A1|2005-03-15|2007-10-18|Brad Hutchings|Hybrid logic/interconnect circuit in a configurable ic|

---

US20070244957A1|2004-11-08|2007-10-18|Jason Redgrave|Configurable IC's with configurable logic circuits that perform adder and/or subtractor operations|

---

US20070241791A1|2004-02-14|2007-10-18|Herman Schmit|Non-Sequentially Configurable IC|

---

US20070241788A1|2004-06-30|2007-10-18|Herman Schmit|VPA Logic Circuits|

---

US20070241786A1|2004-06-30|2007-10-18|Andre Rohe|Configurable Integrated Circuit with Different Connection Schemes|

---

US20070241771A1|2004-02-14|2007-10-18|Herman Schmit|Configurable circuits, IC's, and systems|

---

US20070244960A1|2004-11-08|2007-10-18|Herman Schmit|Configurable IC's with large carry chains|

---

US20070241784A1|2005-03-15|2007-10-18|Brad Hutchings|Configurable ic with interconnect circuits that have select lines driven by user signals|

---

US20070257700A1|2005-03-15|2007-11-08|Andrew Caldwell|Method and apparatus for decomposing functions in a configurable IC|

---

US7295037B2|2004-11-08|2007-11-13|Tabula, Inc.|Configurable IC with routing circuits with offset connections|

---

US20070285125A1|2004-11-08|2007-12-13|Jason Redgrave|Method and Apparatus for Accessing Stored Data in a Reconfigurable IC|

---

US7312631B1|2005-11-01|2007-12-25|Xilinx, Inc.|Structures and methods for avoiding hold time violations in a programmable logic device|

---

US20080059937A1|2004-06-30|2008-03-06|Andre Rohe|Method and apparatus for identifying connections between configurable nodes in a configurable integrated circuit|

---

US20080129333A1|2004-06-30|2008-06-05|Andre Rohe|Configurable Integrated Circuit with Built-in Turns|

---

US7423453B1|2006-01-20|2008-09-09|Advantage Logic, Inc.|Efficient integrated circuit layout scheme to implement a scalable switching network used in interconnection fabric|

---

US20080231318A1|2007-03-20|2008-09-25|Herman Schmit|Configurable ic having a routing fabric with storage elements|

---

US20090146689A1|2007-09-06|2009-06-11|Trevis Chandler|Configuration Context Switcher with a Clocked Storage Element|

---

US20090327987A1|2008-06-26|2009-12-31|Steven Teig|Timing operations in an IC with configurable circuits|

---

US7669097B1|2006-03-27|2010-02-23|Tabula, Inc.|Configurable IC with error detection and correction circuitry|

---

US7679401B1|2005-12-01|2010-03-16|Tabula, Inc.|User registers implemented with routing circuits in a configurable IC|

---

US7694083B1|2006-03-08|2010-04-06|Tabula, Inc.|System and method for providing a virtual memory architecture narrower and deeper than a physical memory architecture|

---

US7705629B1|2008-12-03|2010-04-27|Advantage Logic, Inc.|Permutable switching network with enhanced interconnectivity for multicasting signals|

---

US7714611B1|2008-12-03|2010-05-11|Advantage Logic, Inc.|Permutable switching network with enhanced multicasting signals routing for interconnection fabric|

---

US7765249B1|2005-11-07|2010-07-27|Tabula, Inc.|Use of hybrid interconnect/logic circuits for multiplication|

---

US7797497B1|2006-03-08|2010-09-14|Tabula, Inc.|System and method for providing more logical memory ports than physical memory ports|

---

US7804730B2|2005-03-15|2010-09-28|Tabula, Inc.|Method and apparatus for accessing contents of memory cells|

---

US7818361B1|2005-11-07|2010-10-19|Tabula, Inc.|Method and apparatus for performing two's complement multiplication|

---

US20100327907A1|2009-06-24|2010-12-30|Ting Benjamin S|Enhanced permutable switching network with multicasting signals for interconnection fabric|

---

US7898291B2|2004-12-01|2011-03-01|Tabula, Inc.|Operational time extension|

---

US7930666B1|2006-12-12|2011-04-19|Tabula, Inc.|System and method of providing a memory hierarchy|

---

US7971172B1|2005-11-07|2011-06-28|Tabula, Inc.|IC that efficiently replicates a function to save logic and routing resources|

---

US8112468B1|2007-03-22|2012-02-07|Tabula, Inc.|Method and apparatus for performing an operation with a plurality of sub-operations in a configurable IC|

---

US8463836B1|2005-11-07|2013-06-11|Tabula, Inc.|Performing mathematical and logical operations in multiple sub-cycles|

---

US8863067B1|2008-02-06|2014-10-14|Tabula, Inc.|Sequential delay analysis by placement engines|

---

US8928352B2|2008-09-17|2015-01-06|Tabula, Inc.|Controllable storage elements for an IC|

---

US8941409B2|2011-07-01|2015-01-27|Tabula, Inc.|Configurable storage elements|

---

US9148151B2|2011-07-13|2015-09-29|Altera Corporation|Configurable storage elements|US34363A||1862-02-11||Improvement in machinery for cleaning cotton |

---

US4124899B1|1977-05-23|1987-04-28|||

---

US4870302A|1984-03-12|1989-09-26|Xilinx, Inc.|Configurable electrical circuit having configurable logic elements and configurable interconnects|

---

USRE34363E|1984-03-12|1993-08-31|Xilinx, Inc.|Configurable electrical circuit having configurable logic elements and configurable interconnects|

---

US4642487A|1984-09-26|1987-02-10|Xilinx, Inc.|Special interconnect for configurable logic array|

---

US4706216A|1985-02-27|1987-11-10|Xilinx, Inc.|Configurable logic element|

---

US4821233A|1985-09-19|1989-04-11|Xilinx, Incorporated|5-transistor memory cell with known state on power-up|

---

US4750155A|1985-09-19|1988-06-07|Xilinx, Incorporated|5-Transistor memory cell which can be reliably read and written|

---

US4758745B1|1986-09-19|1994-11-15|Actel Corp|User programmable integrated circuit interconnect architecture and test method|

---

US5157618A|1988-03-10|1992-10-20|Cirrus Logic, Inc.|Programmable tiles|

---

GB8828828D0|1988-12-09|1989-01-18|Pilkington Micro Electronics|Semiconductor integrated circuit|

---

GB8906145D0|1989-03-17|1989-05-04|Algotronix Ltd|Configurable cellular array|

---

US5343406A|1989-07-28|1994-08-30|Xilinx, Inc.|Distributed memory architecture for a configurable logic array and method for using distributed memory|

---

US5260881A|1989-10-30|1993-11-09|Advanced Micro Devices, Inc.|Programmable gate array with improved configurable logic block|

---

US5212652A|1989-08-15|1993-05-18|Advanced Micro Devices, Inc.|Programmable gate array with improved interconnect structure|

---

US5255203A|1989-08-15|1993-10-19|Advanced Micro Devices, Inc.|Interconnect structure for programmable logic device|

---

US5185706A|1989-08-15|1993-02-09|Advanced Micro Devices, Inc.|Programmable gate array with logic cells having configurable output enable|

---

US5231588A|1989-08-15|1993-07-27|Advanced Micro Devices, Inc.|Programmable gate array with logic cells having symmetrical input/output structures|

---

JPH0391932A|1989-09-04|1991-04-17|Canon Inc|Manufacture of semiconductor device|

---

DE4011486A1|1990-04-09|1991-10-10|Inst Stahlbeton Bewehrung Ev|CONCRETE RIBBON STEEL WITH COLD-ROLLED CRANKS AND USE THEREOF|

---

US5267187A|1990-05-10|1993-11-30|Xilinx Inc|Logic structure and circuit for fast carry|

---

US5198705A|1990-05-11|1993-03-30|Actel Corporation|Logic module with configurable combinational and sequential blocks|

---

JPH0447591A|1990-06-14|1992-02-17|Mitsubishi Electric Corp|Semiconductor integrated circuit device|

---

US5073729A|1990-06-22|1991-12-17|Actel Corporation|Segmented routing architecture|

---

US5144166A|1990-11-02|1992-09-01|Concurrent Logic, Inc.|Programmable logic cell and array|

---

US5245227A|1990-11-02|1993-09-14|Atmel Corporation|Versatile programmable logic cell for use in configurable logic arrays|

---

US5079451A|1990-12-13|1992-01-07|Atmel Corporation|Programmable logic device with global and local product terms|

---

US5220213A|1991-03-06|1993-06-15|Quicklogic Corporation|Programmable application specific integrated circuit and logic cell therefor|

---

US5122685A|1991-03-06|1992-06-16|Quicklogic Corporation|Programmable application specific integrated circuit and logic cell therefor|

---

US5313119A|1991-03-18|1994-05-17|Crosspoint Solutions, Inc.|Field programmable gate array|

---

US5241224A|1991-04-25|1993-08-31|Altera Corporation|High-density erasable programmable logic device architecture using multiplexer interconnections|

---

US5338984A|1991-08-29|1994-08-16|National Semiconductor Corp.|Local and express diagonal busses in a configurable logic array|

---

US5260610A|1991-09-03|1993-11-09|Altera Corporation|Programmable logic element interconnections for programmable logic array integrated circuits|

---

US5260611A|1991-09-03|1993-11-09|Altera Corporation|Programmable logic array having local and long distance conductors|

---

US5347519A|1991-12-03|1994-09-13|Crosspoint Solutions Inc.|Preprogramming testing in a field programmable gate array|

---

US5208491A|1992-01-07|1993-05-04|Washington Research Foundation|Field programmable gate array|

---

US5258668A|1992-05-08|1993-11-02|Altera Corporation|Programmable logic array integrated circuits with cascade connections between logic modules|

---

US5384500A|1992-05-15|1995-01-24|Micron Semiconductor, Inc.|Programmable logic device macrocell with an exclusive feedback and an exclusive external input line for a combinatorial mode and accommodating two separate programmable or planes|

---

US5333279A|1992-06-01|1994-07-26|Intel Corporation|Self-timed mesh routing chip with data broadcasting|

---

US5365125A|1992-07-23|1994-11-15|Xilinx, Inc.|Logic cell for field programmable gate array having optional internal feedback and optional cascade|

---

US5645546A|1992-09-14|1997-07-08|Microaire Surgical Instruments|Bone screw and screw driver blade system for oral and maxillofacial surgical procedures|

---

GB9223226D0|1992-11-05|1992-12-16|Algotronix Ltd|Improved configurable cellular array |

---

US5414377A|1992-12-21|1995-05-09|Xilinx, Inc.|Logic block with look-up table for configuration and memory|

---

US5359242A|1993-01-21|1994-10-25|Altera Corporation|Programmable logic with carry-in/carry-out between logic blocks|

---

US5357153A|1993-01-28|1994-10-18|Xilinx, Inc.|Macrocell with product-term cascade and improved flip flop utilization|

---

GB9312674D0|1993-06-18|1993-08-04|Pilkington Micro Electronics|Configurabel logic array|

---

JP2944368B2|1993-07-07|1999-09-06|株式会社東芝|Semiconductor integrated circuit and programmable logic device|

---

US5457410A|1993-08-03|1995-10-10|Btr, Inc.|Architecture and interconnect scheme for programmable logic circuits|

---

US5546596A|1993-08-24|1996-08-13|Intel Corporation|Method and apparatus for integrated local and express routing in a multiprocessor|

---

US5546018A|1993-09-02|1996-08-13|Xilinx, Inc.|Fast carry structure with synchronous input|

---

US5349250A|1993-09-02|1994-09-20|Xilinx, Inc.|Logic structure and circuit for fast carry|

---

US5629886A|1993-09-02|1997-05-13|Xilinx, Inc.|Method and structure for providing fast propagation of a carry signal in a field programmable gate array|

---

US5481206A|1993-09-02|1996-01-02|Xilinx, Inc.|Circuit for fast carry and logic|

---

US5898319A|1993-09-02|1999-04-27|Xilinx, Inc.|Method and structure for providing fast conditional sum in a field programmable gate array|

---

US5455525A|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|

---

KR960002333B1|1993-12-23|1996-02-16|재단법인한국전자통신연구소|Logic integrated circuit module|

---

US5504440A|1994-01-27|1996-04-02|Dyna Logic Corporation|High speed programmable logic architecture|

---

US5682107A|1994-04-01|1997-10-28|Xilinx, Inc.|FPGA architecture with repeatable tiles including routing matrices and logic matrices|

---

US5606266A|1994-11-04|1997-02-25|Altera Corporation|Programmable logic array integrated circuits with enhanced output routing|

---

US5815726A|1994-11-04|1998-09-29|Altera Corporation|Coarse-grained look-up table architecture|

---

US5581199A|1995-01-04|1996-12-03|Xilinx, Inc.|Interconnect architecture for field programmable gate array using variable length conductors|

---

US5537057A|1995-02-14|1996-07-16|Altera Corporation|Programmable logic array device with grouped logic regions and three types of conductors|

---

US5900743A|1995-05-17|1999-05-04|Altera Corporation|Programmable logic array devices with interconnect lines of various lengths|

---

US5543732A|1995-05-17|1996-08-06|Altera Corporation|Programmable logic array devices with interconnect lines of various lengths|

---

US5592106A|1995-05-17|1997-01-07|Altera Corporation|Programmable logic array integrated circuits with interconnection conductors of overlapping extent|

---

US5646546A|1995-06-02|1997-07-08|International Business Machines Corporation|Programmable logic cell having configurable gates and multiplexers|

---

US5583450A|1995-08-18|1996-12-10|Xilinx, Inc.|Sequencer for a time multiplexed programmable logic device|

---

US5848285A|1995-12-26|1998-12-08|Cypress Semiconductor Corporation|Macrocell having a dual purpose input register for use in a logic device|

---

US5635851A|1996-02-02|1997-06-03|Xilinx, Inc.|Read and writable data bus particularly for programmable logic devices|

---

US5677638A|1996-02-02|1997-10-14|Xilinx, Inc.|High speed tristate bus with multiplexers for selecting bus driver|

---

US5724276A|1996-06-17|1998-03-03|Xilinx, Inc.|Logic block structure optimized for sum generation|

---

US5933023A|1996-09-03|1999-08-03|Xilinx, Inc.|FPGA architecture having RAM blocks with programmable word length and width and dedicated address and data lines|

---

US5847580A|1996-10-10|1998-12-08|Xilinx, Inc.|High speed bidirectional bus with multiplexers|

---

US5889413A|1996-11-22|1999-03-30|Xilinx, Inc.|Lookup tables which double as shift registers|

---

US5828230A|1997-01-09|1998-10-27|Xilinx, Inc.|FPGA two turn routing structure with lane changing and minimum diffusion area|

---

US5920202A|1997-02-26|1999-07-06|Xilinx, Inc.|Configurable logic element with ability to evaluate five and six input functions|

---

US5963050A|1997-02-26|1999-10-05|Xilinx, Inc.|Configurable logic element with fast feedback paths|

---

US5914616A|1997-02-26|1999-06-22|Xilinx, Inc.|FPGA repeatable interconnect structure with hierarchical interconnect lines|

---

US5942913A|1997-03-20|1999-08-24|Xilinx, Inc.|FPGA repeatable interconnect structure with bidirectional and unidirectional interconnect lines|

---

US5905385A|1997-04-01|1999-05-18|Advanced Micro Devices, Inc.|Memory bits used to couple look up table inputs to facilitate increased availability to routing resources particularly for variable sized look up tables for a field programmable gate array |

---

US5894426A|1997-10-28|1999-04-13|Winbond Electronics Corporation|Maximum/minimum value determination apparatus|

---

US6154052A|1999-02-04|2000-11-28|Xilinx, Inc.|Combined tristate/carry logic mechanism|US6107822A|1996-04-09|2000-08-22|Altera Corporation|Logic element for a programmable logic integrated circuit|

---

US6204689B1|1997-02-26|2001-03-20|Xilinx, Inc.|Input/output interconnect circuit for FPGAs|

---

US5963050A|1997-02-26|1999-10-05|Xilinx, Inc.|Configurable logic element with fast feedback paths|

---

US6396303B1|1997-02-26|2002-05-28|Xilinx, Inc.|Expandable interconnect structure for FPGAS|

---

US6467009B1|1998-10-14|2002-10-15|Triscend Corporation|Configurable processor system unit|

---

US6124730A|1998-12-15|2000-09-26|Vantis Corporation|Methods for configuring FPGA's having variable grain components for providing time-shared access to interconnect resources|

---

US6407576B1|1999-03-04|2002-06-18|Altera Corporation|Interconnection and input/output resources for programmable logic integrated circuit devices|

---

EP1092268B1|1999-03-04|2008-04-23|Altera Corporation|Interconnection and input/output resources for programmable logic integrated circuit devices|

---

EP1465345A3|1999-03-04|2006-04-12|Altera Corporation|Interconnection resources for programmable logic integrated circuit devices|

---

DE60012639T2|1999-03-04|2005-08-04|Altera Corp., San Jose|CONNECTION RESOURCES FOR PROGRAMMABLE LOGICAL INTEGRATED CIRCUITS|

---

US6633181B1|1999-12-30|2003-10-14|Stretch, Inc.|Multi-scale programmable array|

---

US6721840B1|2000-08-18|2004-04-13|Triscend Corporation|Method and system for interfacing an integrated circuit to synchronous dynamic memory and static memory|

---

US6754760B1|2000-08-21|2004-06-22|Xilinx, Inc.|Programmable interface for a configurable system bus|

---

US7119576B1|2000-09-18|2006-10-10|Altera Corporation|Devices and methods with programmable logic and digital signal processing regions|

---

US7346644B1|2000-09-18|2008-03-18|Altera Corporation|Devices and methods with programmable logic and digital signal processing regions|

---

US7139848B1|2000-12-08|2006-11-21|Xilinx, Inc.|DMA protocol extension for packet-based transfer|

---

US6725364B1|2001-03-08|2004-04-20|Xilinx, Inc.|Configurable processor system|

---

US6895570B2|2001-05-06|2005-05-17|Altera Corporation|System and method for optimizing routing lines in a programmable logic device|

---

US6970014B1|2001-05-06|2005-11-29|Altera Corporation|Routing architecture for a programmable logic device|

---

US6653862B2|2001-05-06|2003-11-25|Altera Corporation|Use of dangling partial lines for interfacing in a PLD|

---

US6605962B2|2001-05-06|2003-08-12|Altera Corporation|PLD architecture for flexible placement of IP function blocks|

---

US6630842B1|2001-05-06|2003-10-07|Altera Corporation|Routing architecture for a programmable logic device|

---

US6466052B1|2001-05-15|2002-10-15|Xilinx, Inc.|Implementing wide multiplexers in an FPGA using a horizontal chain structure|

---

US6476634B1|2002-02-01|2002-11-05|Xilinx, Inc.|ALU implementation in single PLD logic cell|

---

US6838751B2|2002-03-06|2005-01-04|Freescale Semiconductor Inc.|Multi-row leadframe|

---

US6577159B1|2002-04-22|2003-06-10|Nicholas Jesse Macias|Method and apparatus for automatic high-speed bypass routing in a cell matrix self-configurable hardware system|

---

US6670826B1|2002-04-26|2003-12-30|Xilinx, Inc.|Configurable logic block with a storage element clocked by a write strobe pulse|

---

US7112994B2|2002-07-08|2006-09-26|Viciciv Technology|Three dimensional integrated circuits|

---

US7312109B2|2002-07-08|2007-12-25|Viciciv, Inc.|Methods for fabricating fuse programmable three dimensional integrated circuits|

---

US20040004251A1|2002-07-08|2004-01-08|Madurawe Raminda U.|Insulated-gate field-effect thin film transistors|

---

US6747478B2|2002-07-08|2004-06-08|Viciciv|Field programmable gate array with convertibility to application specific integrated circuit|

---

US7064579B2|2002-07-08|2006-06-20|Viciciv Technology|Alterable application specific integrated circuit |

---

US20040018711A1|2002-07-08|2004-01-29|Madurawe Raminda U.|Methods for fabricating three dimensional integrated circuits|

---

US7673273B2|2002-07-08|2010-03-02|Tier Logic, Inc.|MPGA products based on a prototype FPGA|

---

US7064018B2|2002-07-08|2006-06-20|Viciciv Technology|Methods for fabricating three dimensional integrated circuits|

---

US6992503B2|2002-07-08|2006-01-31|Viciciv Technology|Programmable devices with convertibility to customizable devices|

---

US7489164B2|2004-05-17|2009-02-10|Raminda Udaya Madurawe|Multi-port memory devices|

---

US20040019765A1|2002-07-23|2004-01-29|Klein Robert C.|Pipelined reconfigurable dynamic instruction set processor|

---

AU2003252157A1|2002-07-23|2004-02-09|Gatechange Technologies, Inc.|Interconnect structure for electrical devices|

---

US20040111590A1|2002-07-23|2004-06-10|Klein Robert C.|Self-configuring processing element|

---

US6829756B1|2002-09-23|2004-12-07|Xilinx, Inc.|Programmable logic device with time-multiplexed interconnect|

---

US7042756B2|2002-10-18|2006-05-09|Viciciv Technology|Configurable storage device|

---

US7084666B2|2002-10-21|2006-08-01|Viciciv Technology|Programmable interconnect structures|

---

US6946871B1|2002-12-18|2005-09-20|Actel Corporation|Multi-level routing architecture in a field programmable gate array having transmitters and receivers|

---

US6800884B1|2002-12-30|2004-10-05|Actel Corporation|Inter-tile buffer system for a field programmable gate array|

---

US6798240B1|2003-01-24|2004-09-28|Altera Corporation|Logic circuitry with shared lookup table|

---

US6943580B2|2003-02-10|2005-09-13|Altera Corporation|Fracturable lookup table and logic element|

---

US7800401B1|2003-02-10|2010-09-21|Altera Corporation|Fracturable lookup table and logic element|

---

US6888373B2|2003-02-11|2005-05-03|Altera Corporation|Fracturable incomplete look up table for area efficient logic elements|

---

KR101067727B1|2003-02-19|2011-09-28|코닌클리즈케 필립스 일렉트로닉스 엔.브이.|Electronic circuit with array of programmable logic cells|

---

US6930510B2|2003-03-03|2005-08-16|Xilinx, Inc.|FPGA architecture with mixed interconnect resources optimized for fast and low-power routing and methods of utilizing the same|

---

US6943581B1|2003-03-27|2005-09-13|Xilinx, Inc.|Test methodology for direct interconnect with multiple fan-outs|

---

US7000211B2|2003-03-31|2006-02-14|Stretch, Inc.|System and method for efficiently mapping heterogeneous objects onto an array of heterogeneous programmable logic resources|

---

US7613900B2|2003-03-31|2009-11-03|Stretch, Inc.|Systems and methods for selecting input/output configuration in an integrated circuit|

---

US8001266B1|2003-03-31|2011-08-16|Stretch, Inc.|Configuring a multi-processor system|

---

US7581081B2|2003-03-31|2009-08-25|Stretch, Inc.|Systems and methods for software extensible multi-processing|

---

US7590829B2|2003-03-31|2009-09-15|Stretch, Inc.|Extension adapter|

---

US20040242261A1|2003-05-29|2004-12-02|General Dynamics Decision Systems, Inc.|Software-defined radio|

---

US7373642B2|2003-07-29|2008-05-13|Stretch, Inc.|Defining instruction extensions in a standard programming language|

---

US7418575B2|2003-07-29|2008-08-26|Stretch, Inc.|Long instruction word processing with instruction extensions|

---

US6952813B1|2003-07-30|2005-10-04|Xilinx, Inc.|Method and apparatus for selecting programmable interconnects to reduce clock skew|

---

US7129744B2|2003-10-23|2006-10-31|Viciciv Technology|Programmable interconnect structures|

---

US7185035B1|2003-10-23|2007-02-27|Altera Corporation|Arithmetic structures for programmable logic devices|

---

US7243312B1|2003-10-24|2007-07-10|Xilinx, Inc.|Method and apparatus for power optimization during an integrated circuit design process|

---

US7013449B2|2003-10-30|2006-03-14|Hewlett-Packard Development Company, L.P.|Method of designing custom circuit device using scheduling clock cycles|

---

US6982570B1|2003-10-30|2006-01-03|Hewlett-Packard Development Company, L.P.|Reconfigurable device|

---

US7565388B1|2003-11-21|2009-07-21|Altera Corporation|Logic cell supporting addition of three binary words|

---

US7167021B1|2003-11-24|2007-01-23|Altera Corporation|Logic device logic modules having improved arithmetic circuitry|

---

US7317264B2|2003-11-25|2008-01-08|Eaton Corporation|Method and apparatus to independently control contactors in a multiple contactor configuration|

---

US7558812B1|2003-11-26|2009-07-07|Altera Corporation|Structures for LUT-based arithmetic in PLDs|

---

US7030651B2|2003-12-04|2006-04-18|Viciciv Technology|Programmable structured arrays|

---

US7176716B2|2003-12-24|2007-02-13|Viciciv Technology|Look-up table structure with embedded carry logic|

---

US7019557B2|2003-12-24|2006-03-28|Viciciv Technology|Look-up table based logic macro-cells|

---

US7336097B2|2003-12-24|2008-02-26|Viciciv, Inc.|Look-up table structure with embedded carry logic|

---

US7870182B2|2003-12-29|2011-01-11|Xilinx Inc.|Digital signal processing circuit having an adder circuit with carry-outs|

---

US8495122B2|2003-12-29|2013-07-23|Xilinx, Inc.|Programmable device with dynamic DSP architecture|

---

US7467177B2|2003-12-29|2008-12-16|Xilinx, Inc.|Mathematical circuit with dynamic rounding|

---

US7840627B2|2003-12-29|2010-11-23|Xilinx, Inc.|Digital signal processing circuit having input register blocks|

---

US7840630B2|2003-12-29|2010-11-23|Xilinx, Inc.|Arithmetic logic unit circuit|

---

US7853636B2|2003-12-29|2010-12-14|Xilinx, Inc.|Digital signal processing circuit having a pattern detector circuit for convergent rounding|

---

US7865542B2|2003-12-29|2011-01-04|Xilinx, Inc.|Digital signal processing block having a wide multiplexer|

---

US7472155B2|2003-12-29|2008-12-30|Xilinx, Inc.|Programmable logic device with cascading DSP slices|

---

US7480690B2|2003-12-29|2009-01-20|Xilinx, Inc.|Arithmetic circuit with multiplexed addend inputs|

---

US7853632B2|2003-12-29|2010-12-14|Xilinx, Inc.|Architectural floorplan for a digital signal processing circuit|

---

US7853634B2|2003-12-29|2010-12-14|Xilinx, Inc.|Digital signal processing circuit having a SIMD circuit|

---

US7849119B2|2003-12-29|2010-12-07|Xilinx, Inc.|Digital signal processing circuit having a pattern detector circuit|

---

US7882165B2|2003-12-29|2011-02-01|Xilinx, Inc.|Digital signal processing element having an arithmetic logic unit|

---

US7467175B2|2003-12-29|2008-12-16|Xilinx, Inc.|Programmable logic device with pipelined DSP slices|

---

US7844653B2|2003-12-29|2010-11-30|Xilinx, Inc.|Digital signal processing circuit having a pre-adder circuit|

---

US7860915B2|2003-12-29|2010-12-28|Xilinx, Inc.|Digital signal processing circuit having a pattern circuit for determining termination conditions|

---

US7567997B2|2003-12-29|2009-07-28|Xilinx, Inc.|Applications of cascading DSP slices|

---

EP2306331B1|2003-12-29|2018-05-02|Xilinx, Inc.|Integrated circuit with cascading DSP slices|

---

US7176713B2|2004-01-05|2007-02-13|Viciciv Technology|Integrated circuits with RAM and ROM fabrication options|

---

US7193432B1|2004-02-14|2007-03-20|Herman Schmit|VPA logic circuits|

---

US7109752B1|2004-02-14|2006-09-19|Herman Schmit|Configurable circuits, IC's, and systems|

---

KR100564611B1|2004-02-14|2006-03-29|삼성전자주식회사|Damping structure for hard disk drive|

---

US7126381B1|2004-02-14|2006-10-24|Herman Schmit|VPA interconnect circuit|

---

US7622951B2|2004-02-14|2009-11-24|Tabula, Inc.|Via programmable gate array with offset direct connections|

---

US7157933B1|2004-02-14|2007-01-02|Herman Schmit|Configurable circuits, IC's, and systems|

---

US7126373B1|2004-02-14|2006-10-24|Herman Schmit|Configurable logic circuits with commutative properties|

---

US7205791B1|2004-03-12|2007-04-17|Altera Corporation|Bypass-able carry chain in a programmable logic device|

---

US7061268B1|2004-03-15|2006-06-13|Altera Corporation|Initializing a carry chain in a programmable logic device|

---

US7167022B1|2004-03-25|2007-01-23|Altera Corporation|Omnibus logic element including look up table based logic elements|

---

US7698118B2|2004-04-15|2010-04-13|Mentor Graphics Corporation|Logic design modeling and interconnection|

---

JP4576538B2|2004-05-12|2010-11-10|国立大学法人岡山大学|Integrated circuit with multidimensional switch topology|

---

US7635988B2|2007-11-19|2009-12-22|Tier Logic, Inc.|Multi-port thin-film memory devices|

---

US7193438B1|2004-06-30|2007-03-20|Andre Rohe|Configurable integrated circuit with offset connection|

---

US7243329B2|2004-07-02|2007-07-10|Altera Corporation|Application-specific integrated circuit equivalents of programmable logic and associated methods|

---

US7254797B2|2004-09-30|2007-08-07|Rambus Inc.|Input/output cells with localized clock routing|

---

US7342415B2|2004-11-08|2008-03-11|Tabula, Inc.|Configurable IC with interconnect circuits that also perform storage operations|

---

US7428721B2|2004-12-01|2008-09-23|Tabula, Inc.|Operational cycle assignment in a configurable IC|

---

US7496879B2|2004-12-01|2009-02-24|Tabula, Inc.|Concurrent optimization of physical design and operational cycle assignment|

---

US7212030B1|2004-12-31|2007-05-01|Actel Corporation|Field programmable gate array long line routing network|

---

US7358765B2|2005-02-23|2008-04-15|Cswitch Corporation|Dedicated logic cells employing configurable logic and dedicated logic functions|

---

US7224182B1|2005-03-15|2007-05-29|Brad Hutchings|Hybrid configurable circuit for a configurable IC|

---

US7196543B1|2005-06-14|2007-03-27|Xilinx, Inc.|Integrated circuit having a programmable input structure with optional fanout capability|

---

US7215138B1|2005-06-14|2007-05-08|Xilinx, Inc.|Programmable lookup table with dual input and output terminals in shift register mode|

---

US7218140B1|2005-06-14|2007-05-15|Xilinx, Inc.|Integrated circuit having fast interconnect paths between carry chain multiplexers and lookup tables|

---

US7233168B1|2005-06-14|2007-06-19|Xilinx, Inc.|Methods of setting and resetting lookup table memory cells|

---

US7804719B1|2005-06-14|2010-09-28|Xilinx, Inc.|Programmable logic block having reduced output delay during RAM write processes when programmed to function in RAM mode|

---

US7256612B1|2005-06-14|2007-08-14|Xilinx, Inc.|Programmable logic block providing carry chain with programmable initialization values|

---

US7193433B1|2005-06-14|2007-03-20|Xilinx, Inc.|Programmable logic block having lookup table with partial output signal driving carry multiplexer|

---

US7268587B1|2005-06-14|2007-09-11|Xilinx, Inc.|Programmable logic block with carry chains providing lookahead functions of different lengths|

---

US7221186B1|2005-06-14|2007-05-22|Xilinx, Inc.|Efficient tile layout for a programmable logic device|

---

US7276934B1|2005-06-14|2007-10-02|Xilinx, Inc.|Integrated circuit with programmable routing structure including diagonal interconnect lines|

---

US7375552B1|2005-06-14|2008-05-20|Xilinx, Inc.|Programmable logic block with dedicated and selectable lookup table outputs coupled to general interconnect structure|

---

US7274214B1|2005-06-14|2007-09-25|Xilinx, Inc.|Efficient tile layout for a programmable logic device|

---

US7218143B1|2005-06-14|2007-05-15|Xilinx, Inc.|Integrated circuit having fast interconnect paths between memory elements and carry logic|

---

US7279929B1|2005-06-14|2007-10-09|Xilinx, Inc.|Integrated circuit with programmable routing structure including straight and diagonal interconnect lines|

---

US7199610B1|2005-06-14|2007-04-03|Xilinx, Inc.|Integrated circuit interconnect structure having reduced coupling between interconnect lines|

---

US7218139B1|2005-06-14|2007-05-15|Xilinx, Inc.|Programmable integrated circuit providing efficient implementations of arithmetic functions|

---

US7253658B1|2005-06-14|2007-08-07|Xilinx, Inc.|Integrated circuit providing direct access to multi-directional interconnect lines in a general interconnect structure|

---

US7202698B1|2005-06-14|2007-04-10|Xilinx, Inc.|Integrated circuit having a programmable input structure with bounce capability|

---

US7265576B1|2005-06-14|2007-09-04|Xilinx, Inc.|Programmable lookup table with dual input and output terminals in RAM mode|

---

US7202697B1|2005-06-14|2007-04-10|Xilinx, Inc.|Programmable logic block having improved performance when functioning in shift register mode|

---

US7205790B1|2005-06-14|2007-04-17|Xilinx, Inc.|Programmable integrated circuit providing efficient implementations of wide logic functions|

---

US7284228B1|2005-07-19|2007-10-16|Xilinx, Inc.|Methods of using ant colony optimization to pack designs into programmable logic devices|

---

US8620980B1|2005-09-27|2013-12-31|Altera Corporation|Programmable device with specialized multiplier blocks|

---

US7451421B1|2005-11-01|2008-11-11|Xilinx, Inc.|Methods of implementing and modeling interconnect lines at optional boundaries in multi-product programmable IC dies|

---

US7402443B1|2005-11-01|2008-07-22|Xilinx, Inc.|Methods of providing families of integrated circuits with similar dies partially disabled using product selection codes|

---

US7498192B1|2005-11-01|2009-03-03|Xilinx, Inc.|Methods of providing a family of related integrated circuits of different sizes|

---

US7491576B1|2005-11-01|2009-02-17|Xilinx, Inc.|Yield-enhancing methods of providing a family of scaled integrated circuits|

---

US7262633B1|2005-11-11|2007-08-28|Tabula, Inc.|Via programmable gate array with offset bit lines|

---

US8041759B1|2006-02-09|2011-10-18|Altera Corporation|Specialized processing block for programmable logic device|

---

US8301681B1|2006-02-09|2012-10-30|Altera Corporation|Specialized processing block for programmable logic device|

---

US8266198B2|2006-02-09|2012-09-11|Altera Corporation|Specialized processing block for programmable logic device|

---

US8266199B2|2006-02-09|2012-09-11|Altera Corporation|Specialized processing block for programmable logic device|

---

US7486111B2|2006-03-08|2009-02-03|Tier Logic, Inc.|Programmable logic devices comprising time multiplexed programmable interconnect|

---

US7389485B1|2006-03-28|2008-06-17|Xilinx, Inc.|Methods of routing low-power designs in programmable logic devices having heterogeneous routing architectures|

---

US7836117B1|2006-04-07|2010-11-16|Altera Corporation|Specialized processing block for programmable logic device|

---

US7822799B1|2006-06-26|2010-10-26|Altera Corporation|Adder-rounder circuitry for specialized processing block in programmable logic device|

---

US20080024165A1|2006-07-28|2008-01-31|Raminda Udaya Madurawe|Configurable embedded multi-port memory|

---

US8386550B1|2006-09-20|2013-02-26|Altera Corporation|Method for configuring a finite impulse response filter in a programmable logic device|

---

US7930336B2|2006-12-05|2011-04-19|Altera Corporation|Large multiplier for programmable logic device|

---

US8386553B1|2006-12-05|2013-02-26|Altera Corporation|Large multiplier for programmable logic device|

---

US7814137B1|2007-01-09|2010-10-12|Altera Corporation|Combined interpolation and decimation filter for programmable logic device|

---

US8650231B1|2007-01-22|2014-02-11|Altera Corporation|Configuring floating point operations in a programmable device|

---

US7865541B1|2007-01-22|2011-01-04|Altera Corporation|Configuring floating point operations in a programmable logic device|

---

US8645450B1|2007-03-02|2014-02-04|Altera Corporation|Multiplier-accumulator circuitry and methods|

---

US7394287B1|2007-05-21|2008-07-01|Altera Corporation|Programmable logic device having complex logic blocks with improved logic cell functionality|

---

US7459932B1|2007-05-24|2008-12-02|Altera Corporation|Programmable logic device having logic modules with improved register capabilities|

---

US7949699B1|2007-08-30|2011-05-24|Altera Corporation|Implementation of decimation filter in integrated circuit device using ram-based data storage|

---

US8554972B2|2007-09-13|2013-10-08|Friedrich-Alexander-Universitaet-Erlangen-Nuernberg|Logic chip, method and computer program for providing a configuration information for a configurable logic chip|

---

US7902862B2|2007-09-14|2011-03-08|Agate Logic, Inc.|High-bandwidth interconnect network for an integrated circuit|

---

US20090128189A1|2007-11-19|2009-05-21|Raminda Udaya Madurawe|Three dimensional programmable devices|

---

US7812458B2|2007-11-19|2010-10-12|Tier Logic, Inc.|Pad invariant FPGA and ASIC devices|

---

US8643162B2|2007-11-19|2014-02-04|Raminda Udaya Madurawe|Pads and pin-outs in three dimensional integrated circuits|

---

US7602213B2|2007-12-26|2009-10-13|Tier Logic, Inc.|Using programmable latch to implement logic|

---

US7573293B2|2007-12-26|2009-08-11|Tier Logic, Inc.|Programmable logic based latches and shift registers|

---

US7795913B2|2007-12-26|2010-09-14|Tier Logic|Programmable latch based multiplier|

---

US7573294B2|2007-12-26|2009-08-11|Tier Logic, Inc.|Programmable logic based latches and shift registers|

---

US8959137B1|2008-02-20|2015-02-17|Altera Corporation|Implementing large multipliers in a programmable integrated circuit device|

---

US8244789B1|2008-03-14|2012-08-14|Altera Corporation|Normalization of floating point operations in a programmable integrated circuit device|

---

US7701248B2|2008-04-10|2010-04-20|Silicon Storage Technology, Inc.|Storage element for controlling a logic circuit, and a logic device having an array of such storage elements|

---

US8626815B1|2008-07-14|2014-01-07|Altera Corporation|Configuring a programmable integrated circuit device to perform matrix multiplication|

---

US8230375B2|2008-09-14|2012-07-24|Raminda Udaya Madurawe|Automated metal pattern generation for integrated circuits|

---

US8255448B1|2008-10-02|2012-08-28|Altera Corporation|Implementing division in a programmable integrated circuit device|

---

US8307023B1|2008-10-10|2012-11-06|Altera Corporation|DSP block for implementing large multiplier on a programmable integrated circuit device|

---

US8479133B2|2009-01-27|2013-07-02|Xilinx, Inc.|Method of and circuit for implementing a filter in an integrated circuit|

---

US8543635B2|2009-01-27|2013-09-24|Xilinx, Inc.|Digital signal processing block with preadder stage|

---

US8987868B1|2009-02-24|2015-03-24|Xilinx, Inc.|Method and apparatus for programmable heterogeneous integration of stacked semiconductor die|

---

US8805916B2|2009-03-03|2014-08-12|Altera Corporation|Digital signal processing circuitry with redundancy and bidirectional data paths|

---

US8549055B2|2009-03-03|2013-10-01|Altera Corporation|Modular digital signal processing circuitry with optionally usable, dedicated connections between modules of the circuitry|

---

US8645449B1|2009-03-03|2014-02-04|Altera Corporation|Combined floating point adder and subtractor|

---

US8886696B1|2009-03-03|2014-11-11|Altera Corporation|Digital signal processing circuitry with redundancy and ability to support larger multipliers|

---

US8468192B1|2009-03-03|2013-06-18|Altera Corporation|Implementing multipliers in a programmable integrated circuit device|

---

US8706790B1|2009-03-03|2014-04-22|Altera Corporation|Implementing mixed-precision floating-point operations in a programmable integrated circuit device|

---

WO2011008356A2|2009-06-30|2011-01-20|Rambus Inc.|Techniques for adjusting clock signals to compensate for noise|

---

US8650236B1|2009-08-04|2014-02-11|Altera Corporation|High-rate interpolation or decimation filter in integrated circuit device|

---

US7893712B1|2009-09-10|2011-02-22|Xilinx, Inc.|Integrated circuit with a selectable interconnect circuit for low power or high performance operation|

---

US8396914B1|2009-09-11|2013-03-12|Altera Corporation|Matrix decomposition in an integrated circuit device|

---

US8412756B1|2009-09-11|2013-04-02|Altera Corporation|Multi-operand floating point operations in a programmable integrated circuit device|

---

US8341580B2|2009-09-28|2012-12-25|Advantage Logic, Inc.|Modular routing fabric using switching networks|

---

US7948267B1|2010-02-09|2011-05-24|Altera Corporation|Efficient rounding circuits and methods in configurable integrated circuit devices|

---

US8539016B1|2010-02-09|2013-09-17|Altera Corporation|QR decomposition in an integrated circuit device|

---

US8601044B2|2010-03-02|2013-12-03|Altera Corporation|Discrete Fourier Transform in an integrated circuit device|

---

US8458243B1|2010-03-03|2013-06-04|Altera Corporation|Digital signal processing circuit blocks with support for systolic finite-impulse-response digital filtering|

---

US8484265B1|2010-03-04|2013-07-09|Altera Corporation|Angular range reduction in an integrated circuit device|

---

US8510354B1|2010-03-12|2013-08-13|Altera Corporation|Calculation of trigonometric functions in an integrated circuit device|

---

US8539014B2|2010-03-25|2013-09-17|Altera Corporation|Solving linear matrices in an integrated circuit device|

---

US9015023B2|2010-05-05|2015-04-21|Xilinx, Inc.|Device specific configuration of operating voltage|

---

US8862650B2|2010-06-25|2014-10-14|Altera Corporation|Calculation of trigonometric functions in an integrated circuit device|

---

US8589463B2|2010-06-25|2013-11-19|Altera Corporation|Calculation of trigonometric functions in an integrated circuit device|

---

US8577951B1|2010-08-19|2013-11-05|Altera Corporation|Matrix operations in an integrated circuit device|

---

US8890567B1|2010-09-30|2014-11-18|Altera Corporation|High speed testing of integrated circuits including resistive elements|

---

US8159265B1|2010-11-16|2012-04-17|Raminda Udaya Madurawe|Memory for metal configurable integrated circuits|

---

US8159268B1|2010-11-16|2012-04-17|Raminda Udaya Madurawe|Interconnect structures for metal configurable integrated circuits|

---

US8159266B1|2010-11-16|2012-04-17|Raminda Udaya Madurawe|Metal configurable integrated circuits|

---

JP5701054B2|2010-12-28|2015-04-15|国立大学法人 熊本大学|Reconfigurable logic block, programmable logic circuit device using the same, and technology mapping method|

---

US8645451B2|2011-03-10|2014-02-04|Altera Corporation|Double-clocked specialized processing block in an integrated circuit device|

---

US9600278B1|2011-05-09|2017-03-21|Altera Corporation|Programmable device using fixed and configurable logic to implement recursive trees|

---

US8812576B1|2011-09-12|2014-08-19|Altera Corporation|QR decomposition in an integrated circuit device|

---

US9053045B1|2011-09-16|2015-06-09|Altera Corporation|Computing floating-point polynomials in an integrated circuit device|

---

US8949298B1|2011-09-16|2015-02-03|Altera Corporation|Computing floating-point polynomials in an integrated circuit device|

---

US8762443B1|2011-11-15|2014-06-24|Altera Corporation|Matrix operations in an integrated circuit device|

---

US8522185B1|2012-02-06|2013-08-27|Xilinx, Inc.|Method for placement and routing of a circuit design|

---

US8543634B1|2012-03-30|2013-09-24|Altera Corporation|Specialized processing block for programmable integrated circuit device|

---

US9166598B1|2012-05-08|2015-10-20|Altera Corporation|Routing and programming for resistive switch arrays|

---

US9098332B1|2012-06-01|2015-08-04|Altera Corporation|Specialized processing block with fixed- and floating-point structures|

---

US8957512B2|2012-06-19|2015-02-17|Xilinx, Inc.|Oversized interposer|

---

US8869088B1|2012-06-27|2014-10-21|Xilinx, Inc.|Oversized interposer formed from a multi-pattern region mask|

---

US8996600B1|2012-08-03|2015-03-31|Altera Corporation|Specialized processing block for implementing floating-point multiplier with subnormal operation support|

---

US9026872B2|2012-08-16|2015-05-05|Xilinx, Inc.|Flexible sized die for use in multi-die integrated circuit|

---

US9490811B2|2012-10-04|2016-11-08|Efinix, Inc.|Fine grain programmable gate architecture with hybrid logic/routing element and direct-drive routing|

---

US9525419B2|2012-10-08|2016-12-20|Efinix, Inc.|Heterogeneous segmented and direct routing architecture for field programmable gate array|

---

US9207909B1|2012-11-26|2015-12-08|Altera Corporation|Polynomial calculations optimized for programmable integrated circuit device structures|

---

CN105027446B|2013-03-01|2019-06-21|軸子研究株式会社|Data processing equipment and its control method|

---

US9189200B1|2013-03-14|2015-11-17|Altera Corporation|Multiple-precision processing block in a programmable integrated circuit device|

---

US9000490B2|2013-04-19|2015-04-07|Xilinx, Inc.|Semiconductor package having IC dice and voltage tuners|

---

US9348795B1|2013-07-03|2016-05-24|Altera Corporation|Programmable device using fixed and configurable logic to implement floating-point rounding|

---

US9547034B2|2013-07-03|2017-01-17|Xilinx, Inc.|Monolithic integrated circuit die having modular die regions stitched together|

---

US9379687B1|2014-01-14|2016-06-28|Altera Corporation|Pipelined systolic finite impulse response filter|

---

US9915869B1|2014-07-01|2018-03-13|Xilinx, Inc.|Single mask set used for interposer fabrication of multiple products|

---

US9372956B1|2014-11-10|2016-06-21|Xilinx, Inc.|Increased usable programmable device dice|

---

US9684488B2|2015-03-26|2017-06-20|Altera Corporation|Combined adder and pre-adder for high-radix multiplier circuit|

---

US10942706B2|2017-05-05|2021-03-09|Intel Corporation|Implementation of floating-point trigonometric functions in an integrated circuit device|

法律状态:
2001-08-30| STCF| Information on status: patent grant|Free format text: PATENTED CASE |

---

2005-03-21| FPAY| Fee payment|Year of fee payment: 4 |

---

2005-03-21| SULP| Surcharge for late payment|

---

2009-03-16| FPAY| Fee payment|Year of fee payment: 8 |

---

2012-10-02| FPAY| Fee payment|Year of fee payment: 12 |

优先权:

---

申请号 | 申请日 | 专利标题

---

US08/806,997|US5914616A|1997-02-26|1997-02-26|FPGA repeatable interconnect structure with hierarchical interconnect lines|

---

US08/823,265|US5963050A|1997-02-26|1997-03-24|Configurable logic element with fast feedback paths|

---

US09/311,782|US6107827A|1997-02-26|1999-05-13|FPGA CLE with two independent carry chains|

---

US09/574,741|US6204690B1|1997-02-26|2000-05-18|FPGA architecture with offset interconnect lines|

---

US09/759,051|US6292022B2|1997-02-26|2001-01-11|Interconnect structure for a programmable logic device|US09/759,051| US6292022B2|1997-02-26|2001-01-11|Interconnect structure for a programmable logic device|

---

US09/929,977| US6448808B2|1997-02-26|2001-08-15|Interconnect structure for a programmable logic device|