US20070241785A1 - Configurable ic's with logic resources with offset connections - Google Patents

Abstract

A configurable integrated circuit (“IC”) that includes several configurable tiles arranged in a tile arrangement. Each configurable tile has a set of configurable logic circuits and a set of configurable routing circuits for routing signals between configurable logic circuits. In some embodiments, at least a first logic circuit of a first tile has at least one direct connection with a second circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.

Description

FIELD OF THE INVENTION
• The present invention is directed towards configurable IC's with logic resources with offset connections.
BACKGROUND OF THE INVENTION
• The use of configurable integrated circuits (“IC's”) has dramatically increased in recent years. One example of a configurable IC is a field programmable gate array (“FPGA”). An FPGA is a field programmable IC that usually has logic circuits, interconnect circuits, and input/output (I/O) circuits. The logic circuits (also called logic blocks) are typically arranged as an internal array of circuits. These logic circuits are connected together through numerous interconnect circuits (also called interconnects). The logic and interconnect circuits are typically surrounded by the I/O circuits.
• FIG. 1 illustrates an example of aconfigurable logic circuit100. This logic circuit can be configured to perform a number of different functions. As shown inFIG. 1, thelogic circuit100 receives a set ofinput data105 and a set ofconfiguration data110. The configuration data set can be stored in a set ofSRAM cells115. From the set of functions that thelogic circuit100 can perform, the configuration data set specifies a particular function that this circuit is to perform on the input data set. Once the logic circuit performs its function on the input data set, it provides the output of this function on a set ofoutput lines120. Thelogic circuit100 is said to be configurable, as the configuration data set “configures” the logic circuit to perform a particular function, and this configuration data set can be modified by writing new data in the SRAM cells. Multiplexers and look-up tables are two examples of configurable logic circuits.
• FIG. 2 illustrates an example of aconfigurable interconnect circuit200. Thisinterconnect circuit200 connects a set ofinput data205 to a set ofoutput data210. This circuit receivesconfiguration data bits215 that are stored in a set ofSRAM cells220. The configuration bits specify how the interconnect circuit should connect the input data set to the output data set. Theinterconnect circuit200 is said to be configurable, as the configuration data set “configures” the interconnect circuit to use a particular connection scheme that connects the input data set to the output data set in a desired manner. Moreover, this configuration data set can be modified by writing new data in the SRAM cells. Multiplexers are one example of interconnect circuits.
• FIG. 3 illustrates a portion of a prior art configurable IC300. As shown in this figure, the IC300 includes an array ofconfigurable logic circuits305 and configurable interconnect circuits310. The IC300 has two types ofinterconnect circuits310aand310b.Interconnect circuits310aconnectinterconnect circuits310bandlogic circuits305, while interconnectcircuits310bconnectinterconnect circuits310atoother interconnect circuits310a. In some cases, the IC300 includes hundreds or thousands oflogic circuits305 and interconnect circuits310.
• In some configurable IC architectures, aninterconnect circuit310bcan connect to interconnectcircuits310bthat are several columns or several rows away from it in the array.FIG. 4 illustrates several such connections in a priorconfigurable IC architecture400. In thearchitecture400, eachlogic circuit305 forms a configurablecomputational tile405 in conjunction with two neighboringinterconnect circuits310aand one neighboringinterconnect circuit310b. In each particular tile, eachinterconnect circuit310acan receive inputs from theinterconnect circuit310bin the tile and supply a sub-set of the received input signals (e.g., one input signal) to thelogic circuit305 of the tile.
• Theinterconnect circuits310bin each particular tile serve as switchboxes that connect toother interconnect circuits310bthrough interveninginterconnect circuits310a. As shown inFIG. 4, theseswitchboxes310bcan also connect toother switchboxes310bthat are two or more rows or columns away but in the same column or row. For instance, each switchbox can connect to switchboxes that are one, two, three and six rows above and below it, and to switchboxes that are one, two, three, and six columns to its right and left.
• In the architecture ofFIG. 4, aparticular logic circuit305 connects to logic circuits that are in the four tiles that are diagonally adjacent to the particular logic circuit's tile, through fourconnection boxes310ain these tiles. For instance,FIG. 4 illustrates that thelogic circuit305 intile405aconnects to thelogic circuits305 intiles405b-e through aconnection box310ain these tiles.
• The advantage of the connection architecture illustrated inFIG. 4 is that it allows one computation tile to connect to another computational tile that is not a neighboring tile. On the other hand, this architecture requires the use of multiple connections to connect two tiles that are not diagonally adjacent and that are in two different rows and columns. This requirement makes the connection architecture illustrated inFIG. 4 inefficient and expensive as each connection requires the use of transistor switching logic.
• Also, the connection architecture illustrated inFIG. 4 employs the same set of long connection schemes for each tile. Hence, as shown inFIG. 5, this architecture can result in a loop between twotiles505 and510 in the same column, or twotiles515 and520 in the same row. Such cycles are undesirable as they come at the expense of reachability of other tiles. The uniform connection architecture ofFIG. 4 is also inefficient as it provides more ways than necessary for reaching one tile from another tile. This redundancy is illustrated inFIG. 5, which illustrates that thetile525 can connect totile530 through two different sets of connections, one that goes throughtile535 and one that goes throughtile540. This redundancy is undesirable as it comes at the expense of reachability of other tiles.
• Therefore, there is a need in the art for a configurable IC that has a wiring architecture that increases the interconnectivity between its configurable circuits.
SUMMARY OF THE INVENTION
• Some embodiments provide a configurable integrated circuit (“IC”) that includes several configurable tiles arranged in a tile arrangement. Each configurable tile has a set of configurable logic circuits and a set of configurable routing circuits for routing signals between configurable logic circuits.
• In some embodiments, at least a first logic circuit of a first tile has at least one direct connection with a second circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.
• Also, in some embodiments, each particular tile further has a set of configurable input-select circuits for receiving inputs and configurably supplying a sub-set of the received inputs to the configurable logic circuits in the particular tile. At least a first input select circuit of a first tile has at least one direct connection with a second circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement, where the direction connection is for supplying a signal to the first input select circuit.
• In some embodiments, several logic circuits of the first tile have at least one direct connection with at least one circuit of another tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement. In some embodiments, several input select circuits of the first tile have at least one direct connection with at least one circuit of another tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.
BRIEF DESCRIPTION OF THE DRAWINGS
• The novel features of the invention are set forth in the appended claims. However, for purpose of explanation, several embodiments of the invention are set forth in the following figures.
• FIG. 1 illustrates an example of a configurable logic circuit.
• FIG. 2 illustrates an example of a configurable interconnect circuit.
• FIG. 3 illustrates a portion of a prior art configurable IC.
• FIG. 4 illustrates several connections in a prior configurable IC architecture.
• FIG. 5 illustrates an IC architecture that results in a loop between two tiles in the same column, or two tiles in the same row.
• FIG. 6 illustrates an example of a direct connection where all the wire segments that establish a direct connection are on the same layer.
• FIG. 7 illustrates an example of a direct connection where the connecting wire segments and the terminals of the connected circuits are all on the same layer.
• FIG. 8 illustrates an example of a direct connection where the set of wire segments that establish the direct connection between two circuits are on several wiring layers.
• FIG. 9 illustrates an example of a direct connection between two circuits established by one or more diagonal wire segments possibly in conjunction with one or more Manhattan (i.e., horizontal or vertical) segments.
• FIG. 10 illustrates an example of using one buffer circuit in the direct connection between circuits.
• FIG. 11 illustrates an example of using two buffer circuits in the direct connection between circuits.
• FIG. 12 illustrates an example of a configurable logic circuit that can perform a set of functions.
• FIG. 13 illustrates an example of a configurable interconnect circuit.
• FIG. 14 illustrates an example of a sub-cycle reconfigurable IC.
• FIG. 15 illustrates an example of a reconfigurable logic circuit.
• FIG. 16 illustrates an example of a reconfigurable interconnect circuit.
• FIG. 17 illustrates an IC architecture that is formed by numerous rectangular configurable tiles that are arranged in an array with multiple rows and columns.
• FIG. 18 illustrates a first input select multiplexer connected to four neighboring LUT's, two offset LUT's, and two offset routing multiplexers.
• FIG. 19 illustrates a second input select multiplexer connected to four neighboring offset LUT's, two other offset LUT's, and two offset routing multiplexers.
• FIG. 20 illustrates a third input select multiplexer connected to eight neighboring offset LUT's.
• FIG. 21 illustrates a first routing multiplexer connected to four neighboring LUT's and to four horizontally or vertically aligned routing multiplexers.
• FIG. 22 illustrates a second routing multiplexer connects to the four LUT's and to four horizontally or vertically aligned routing multiplexers.
• FIG. 23 illustrates an example of an architecture that is asymmetric with respect to the inputs of the routing interconnects.
• FIG. 24 illustrates a set of Boolean gates that compute two functions based on a set of inputs.
• FIG. 25 illustrates the design ofFIG. 24 after its gates have been placed into four groups.
• FIG. 26 illustrates another representation of the design ofFIG. 24.
• FIG. 27 illustrates a circuit representation of an interconnect/storage circuit that can be used to implement the routing multiplexer of some embodiments.
• FIG. 28 illustrates an HUMUX that includes two two-to-one multiplexers, a four-to-one multiplexer, a set of input terminals, an output terminal, and a set of select terminals.
• FIG. 29 illustrates a portion of the architecture of a configurable IC.
• FIG. 30 illustrates a portion of the actual physical architecture of the configurable IC.
• FIG. 31 illustrates an aligned tile layout, which is formed by four tiles that are aligned in the physical architecture.
• FIG. 32 illustrates a logic carry block (LCB) that is formed by a three-input LUT and its associated carry logic circuit.
• FIG. 33 illustrates an alternative carry-signal flow through four, aligned LCB's.
• FIG. 34 illustrates two fast nibble wide adders/subtractors that are on the same topological row ganged to form a fast byte-wide adder/subtractor.
• FIG. 35 illustrates an aligned layout that includes one common carry chain that is shared among the four logic circuits in the tile layout.
• FIG. 36 illustrates a bypass circuitry to bypass the shared carry logic to further speed the carry logic circuitry for larger adders/subtractors.
• FIG. 37 illustrates an example of a three-input LUT.
• FIG. 38 illustrates a three-input LUT that is an optimized version of the LUT ofFIG. 37.
• FIG. 39 illustrates a CPL-implementation of a four-stage Manchester carry chain that can serve as the shared carry logic ofFIG. 36.
• FIG. 40 illustrates a tile group that includes two carry chains, a left-to-right carry chain and a right-to-left carry chain.
• FIG. 41 illustrates a tile layout that includes two Manchester carry logics, two routing multiplexers, and two sets of carry in and out signals.
• FIG. 42 illustrates one manner of embedding a memory in the layout of the tile group ofFIG. 40.
• FIG. 43 illustrates a physical layout for embedding a memory in an aligned tile group, which is formed by four tiles that are aligned with each other in a manner similar to the aligned tile groups ofFIGS. 31 and 41.
• FIG. 44 illustrates an architecture that includes address and data signals for a memory that come from several groups of tiles.
• FIG. 45 illustrates a manner for establishing the dual-ported architecture ofFIG. 43.
• FIG. 46 illustrates a portion of a configurable IC.
• FIG. 47 illustrates a more detailed example of a configuration data pool for the configurable IC.
• FIG. 48 illustrates a system on chip (“SoC”) implementation of a configurable IC.
• FIG. 49 illustrates a system in package (“SiP”) implementation for a configurable IC.
• FIG. 50 illustrates a more detailed example of a computing system that has a configurable IC.
DETAILED DESCRIPTION OF THE INVENTION
• In the following description, numerous details are set forth for purpose of explanation. However, one of ordinary skill in the art will realize that the invention may be practiced without the use of these specific details. For instance, not all embodiments of the invention need to be practiced with the specific number of bits and/or specific devices (e.g., multiplexers) referred to below. In other instances, well-known structures and devices are shown in block diagram form in order not to obscure the description of the invention with unnecessary detail.
• Some embodiments of the invention provide architectures for configurable IC's that have configurable computational units (e.g., configurable logic circuits) and configurable routing circuits for configurably routing signals between the configurable computational units. For instance, some embodiments provide a configurable IC that includes numerous configurable computational tiles (e.g., hundreds, thousands, hundreds of thousands, etc. of tiles) that are laid out on the IC according to a particular arrangement. In some embodiments, the configurable computational tiles include configurable logic circuits and configurable interconnect circuits. In other embodiments, the only configurable circuits in the configurable computational tiles are configurable logic circuits or configurable interconnect circuits.
• The computational tiles in some embodiments are arranged in numerous rows and columns that form a tile array. Also, the tile arrangement in some embodiments result in one or more sets of the configurable circuits (e.g., the configurable logic circuits and/or configurable interconnect circuits) being arranged in an array with several aligned rows and columns. Alternatively, some embodiments might organize the configurable circuits in an arrangement that is not an array.
• Accordingly, instead of referring to configurable circuit arrays or configurable tile arrays, the discussion below refers to configurable circuit arrangements and configurable tile arrangements. Some arrangements may have configurable circuits or tiles arranged in one or more arrays, while other arrangements may not have the configurable circuits or tiles arranged in an array. In the tile or circuit arrangement, some embodiments intersperse several other circuits, such as memory blocks, processors, macro blocks, IP blocks, SERDES controllers, clock management units, etc. Alternatively, some embodiments arrange some of these other circuits (e.g., memory blocks) within the tile structure.
• Each computation tile in some embodiments includes a set of configurable logic circuits and a set of configurable routing circuits (also called configurable routing fabric or resources). In some embodiments, the set of configurable logic circuits in each computational tile includes a set of input select interconnect circuits associated with the set of configurable logic circuits.
• In some embodiments, each routing interconnect circuit can receive several input signals and distribute output signals to several different types of circuits, such as input select interconnect(s) of the same computational tile, or routing and input-select interconnects of other tiles. In some embodiments, at least one routing interconnect of a particular computational tile can receive signals from, and supply signals to, only circuits outside of the particular tile. In some embodiments, one routing interconnect in a particular computational tile is not connected to any other circuit in its own tile or in any tile that neighbors its own tile. Also, routing interconnects can have fan out greater than one in some embodiments.
• Alternatively, in some embodiments, the input select interconnects of a computational tile supply their output signals to only the logic circuits of the particular tile. Specifically, each input select interconnect of these embodiments receives input signals for at least one logic circuit and supplies a sub-set of the received inputs to the particular logic circuit set. In some of these embodiments, each input select interconnect of a computational tile provides its output to only one logic circuit (i.e., each such input select interconnect has a fan out of one).
• In some embodiments, one or more input select interconnects of a particular computational tile directly receives input from one or more circuits outside of the particular tile. As further described below, a direct connection between two circuits is an electrical connection between the two circuits that is achieved by (1) a set of wire segments that traverse through a set of the wiring layers of the IC, and (2) a set of vias when two or more wiring layers are involved. In some embodiments, a direct connection between two circuits might also include a set of buffer circuits.
• Through its direct connections with circuits outside of its particular computational tile, a particular computational tile's input select interconnects can receive input signals from the circuits outside of the particular tile, and pass a set of these received signals to a logic circuit in the particular computational tile. In some of these embodiments, the particular computational tile's input select interconnects have direct connections with circuits in tiles that are several tiles away from the particular tile. In some of these embodiments, one or more of these other tiles are not vertically or horizontally aligned with the particular computational tile in the tile arrangement. In other words, some embodiments have several long direct offset connections for connecting the inputs of some input select interconnects with circuits that are in computational tiles that are offset from the particular computational tile by several rows and/or columns.
• Some embodiments also have several offset connections between interconnects in different computational tiles. For instance, in some embodiments, the output of a routing interconnect in a particular computational tile can be supplied through an offset connection to the input of the routing interconnect of another computational tile. Such an offset connect can also be used to provide the output of a routing interconnect in one computational tile to the input select interconnect in another computational tile. Some embodiments use long offset connections to connect two interconnects that are neither in neighboring computational tiles, nor in vertically or horizontally aligned computational tiles. Some embodiments also use a long offset connection to provide the output of logic circuits to circuits that are in offset computational tiles that do not neighbor the computational tiles of the logic circuits.
• The use of direct offset connections in the configurable IC of some embodiments increases the interconnectivity between the circuits of the configurable IC. In addition to computational tiles, some embodiments include other types of tiles (e.g., tiles that embed memory arrays) that do not include some or all of the circuits of a computational tile. In some embodiments, these other tiles connect to each other and/or to computational tiles in the same manner as was described above for connections between computational tiles. The configurable IC of some embodiments is a reconfigurable IC. In some of these embodiments, the reconfigurable IC is a sub-cycle reconfigurable IC.
• Several more detailed embodiments of the invention are described in Sections II-X of the detailed description. However, before this description, several terms and concepts are discussed in Section I.
• I. Terms and Concepts
• A. Direct Connections Between Circuits
• Several figures below illustrate several direct connections between circuits in a configurable circuit arrangement. A direct connection between two circuits in an arrangement is an electrical connection between the two circuits that is achieved by (1) a set of wire segments that traverse through a set of the wiring layers of the IC, and (2) a set of vias when two or more wiring layers are involved.
• FIGS. 6-9 illustrate several examples of direct connections between two circuits. These examples illustrate actual geometric realization of the direct connections.FIG. 6 illustrates a case where all the wire segments that establish a direct connection are on the same layer. Specifically, this figure illustrates fourwire segments620,625,630, and635 that establish the direct connection betweencircuits605 and610, which are offset in the circuit arrangement of a configurable IC. These four segments might be on a layer (e.g., the second wiring layer) that is different from the layer (e.g., the first wiring layer) that has the input/output terminals615 and640 of thecircuits605 and610. Hence, in these cases, the direct connection between thecircuits605 and610 also require a set ofvias645 and650 to connect thewire segments620 and635 to theterminals615 and640.
• FIG. 7 illustrates an example were the connectingwire segments715 and the terminals of theconnected circuits705 and710 are all on the same layer. Alternatively,FIG. 8 illustrates a case where the set of wire segments that establish a direct connection between two circuits are on several wiring layers. In this example, a direct connection is established between the twocircuits805 and810 by (1) a vertical segment825 (e.g., a segment in the y-direction on layer 2) that connects to a horizontal terminal815 (e.g., a segment in the x-direction on layer 1) of thecircuit805 through a viaconnection820, and (2) a horizontal segment835 (on layer 3) that connects to a vertical terminal845 (on layer 1) of thecircuit810 through a stacked viaconnection840. Thehorizontal segment835 also connects to thevertical segment825 through a viaconnection830.
• When the IC uses a wiring model that allows occasional or systematic diagonal wiring, a direct connection between two circuits can be established by one or more diagonal wire segments possibly in conjunction with one or more Manhattan (i.e., horizontal or vertical) segments.FIG. 9 illustrates an example of such a direct connection. Specifically, this figure illustrates a 60° diagonal segment925 (e.g., on a third wiring layer) that connects to the vertical terminal915 (on layer 1) ofcircuit905 and the vertical terminal935 (on layer 1) ofcircuit910 through stacked viaconnections920 and930.
• The direct connection illustrated inFIGS. 7-9 are examples of built-in turns used by some embodiments of the invention. Built-in turns allow two offset circuits to be connected by relying on wiring architecture that reduces the number of interconnect circuits necessary for establishing the connection between the two circuits. Built-in turns are further described in U.S. patent application Ser. No. 10/882,845, entitled “Configurable Integrated Circuit with Built-In Turns”, and filed Jun. 30, 2004.
• In some embodiments, a direct connection between two circuits in an arrangement might also include a set of buffer circuits in some cases. In other words, two circuits are connected in some embodiments by a set of wire segments that possibly traverse through a set of buffer circuits and a set of vias. Buffer circuits are not interconnect circuits or configurable logic circuits. In some embodiments, buffer circuits are part of some or all connections. Buffer circuits might be used to achieve one or more objectives (e.g., maintain the signal strength, reduce noise, alter signal delay, etc.) along the wire segments that establish the direct connections. Inverting buffer circuits may also allow an IC design to reconfigure logic circuits less frequently and/or use fewer types of logic circuits. In some embodiments, buffer circuits are formed by one or more inverters (e.g., two or more inverters that are connected in series).FIGS. 10 and 11 illustrate examples of using one or twobuffer circuits1005 and1105 in the direct connection betweencircuits605 and610 ofFIG. 6.
• Alternatively, the intermediate buffer circuits between the logic and/or interconnect circuits can be viewed as a part of the devices illustrated in these figures. For instance, the inverters that can be placed after thedevices605 and610 can be viewed as being part of these devices. Some embodiments use such inverters in order to allow an IC design to reconfigure logic circuits less frequently and/or use fewer types of logic circuits.
• Several figures below “topologically” illustrate several direct connections between circuits in an arrangement. A topological illustration is an illustration that is only meant to show a direct connection between two circuits without specifying a particular geometric layout for the wire segments that establish the direct connection.
• B. Configurable and Reconfigurable IC's
• A configurable IC is an IC that has configurable circuits. In some embodiments, a configurable IC includes configurable computational circuits (e.g., configurable logic circuits) and configurable routing circuits for routing the signals to and from the configurable computation units. In addition to configurable circuits, a configurable IC also typically includes non-configurable circuits (e.g., non-configurable logic circuits, interconnect circuits, memories, etc.).
• A configurable circuit is a circuit that can “configurably” perform a set of operations. Specifically, a configurable circuit receives “configuration data” that specifies the operation that the configurable circuit has to perform in the set of operations that it can perform. In some embodiments, configuration data is generated outside of the configurable IC. In these embodiments, a set of software tools typically converts a high-level IC design (e.g., a circuit representation or a hardware description language design) into a set of configuration data that can configure the configurable IC (or more accurately, the configurable IC's configurable circuits) to implement the IC design.
• Examples of configurable circuits include configurable interconnect circuits and configurable logic circuits. A logic circuit is a circuit that can perform a function on a set of input data that it receives. A configurable logic circuit is a logic circuit that can be configured to perform different functions on its input data set.
• FIG. 12 illustrates an example of aconfigurable logic circuit1200 that can perform a set of functions. As shown in this figure, thelogic circuit1200 has a set ofinput terminals1205, a set ofoutput terminals1210, and a set ofconfiguration terminals1215. Thelogic circuit1200 receives a set of configuration data along itsconfiguration terminals1215. Based on the configuration data, the logic circuit performs a particular function within its set of functions on the input data that it receives along itsinput terminals1205. The logic circuit then outputs the result of this function as a set of output data along itsoutput terminal set1210. Thelogic circuit1200 is said to be configurable as the configuration data set “configures” the logic circuit to perform a particular function.
• A configurable interconnect circuit is a circuit that can configurably connect an input set to an output set in a variety of ways.FIG. 13 illustrates an example of aconfigurable interconnect circuit1300. Thisinterconnect circuit1300 connects a set ofinput terminals1305 to a set ofoutput terminals1310, based on a set ofconfiguration data1315 that the interconnect circuit receives. In other words, the configuration data specify how the interconnect circuit should connect the input terminal set1305 to theoutput terminal set1310. Theinterconnect circuit1300 is said to be configurable as the configuration data set “configures” the interconnect circuit to use a particular connection scheme that connects the input terminal set to the output terminal set in a desired manner.
• An interconnect circuit can connect two terminals or pass a signal from one terminal to another by establishing an electrical path between the terminals. Alternatively, an interconnect circuit can establish a connection or pass a signal between two terminals by having the value of a signal that appears at one terminal appear at the other terminal. In connecting two terminals or passing a signal between two terminals, an interconnect circuit in some embodiments might invert the signal (i.e., might have the signal appearing at one terminal inverted by the time it appears at the other terminal). In other words, the interconnect circuit of some embodiments implements a logic inversion operation in conjunction to its connection operation. Other embodiments, however, do not build such an inversion operation in some or all of their interconnect circuits.
• Reconfigurable IC's are one type of configurable IC's. Specifically, reconfigurable IC's are configurable IC's that can reconfigure during runtime.FIG. 14 conceptually illustrates an example of a sub-cycle reconfigurable IC (i.e., an IC that is reconfigurable on a sub-cycle basis). In this example, the sub-cycle reconfigurable IC implements anIC design1405 that operates at a clock speed of X MHz. Typically, an IC design is initially specified in a hardware description language (HDL), and a synthesis operation is used to convert this HDL representation into a circuit representation. After the synthesis operation, the IC design includes numerous electronic circuits, which are referred to below as “components.” As further illustratedFIG. 14, the operations performed by the components in theIC design1405 can be partitioned into four sets of operations1410-1425, with each set of operations being performed at a clock speed of X MHz.
• FIG. 14 then illustrates that these four sets of operations1410-1425 can be performed by one sub-cyclereconfigurable IC1430 that operates at 4X MHz. In some embodiments, four cycles of the 4X MHz clock correspond to four sub-cycles within a cycle of the X MHz clock. Accordingly, this figure illustrates thereconfigurable IC1430 reconfiguring four times during four cycles of the 4X MHz clock (i.e., during four sub-cycles of the X MHz clock). During each of these reconfigurations (i.e., during each sub-cycle), thereconfigurable IC1430 performs one of the identified four sets of operations. In other words, the faster operational speed of thereconfigurable IC1430 allows this IC to reconfigure four times during each cycle of the X MHz clock, in order to perform the four sets of operations sequentially at a 4X MHz rate instead of performing the four sets of operations in parallel at an X MHz rate.
• A reconfigurable IC typically includes reconfigurable logic circuits and/or reconfigurable interconnect circuits, where the reconfigurable logic and/or interconnect circuits are configurable logic and/or interconnect circuits that can “reconfigure” more than once at runtime. A configurable logic or interconnect circuit reconfigures when it bases its operation on a different set of configuration data.
• FIG. 15 illustrates an example of areconfigurable logic circuit1500. This logic circuit includes acore logic circuit1505 that can perform a variety of functions on a set ofinput data1510 that it receives. Thecore logic circuit1505 also receives a set of fourconfiguration data bits1515 through aswitching circuit1520, which in this case is formed by four four-to-one multiplexers1540. The switching circuit receives a larger set of sixteenconfiguration data bits1525 that, in some cases, are stored in a set of storage elements1530 (e.g., a set of memory cells, such as SRAM cells). This switching circuit is controlled by a two-bit reconfiguration signal φ through twoselect lines1555. Whenever the reconfiguration signal changes, the switching circuit supplies a different set of four configuration data bits to thecore logic circuit1505. The configuration data bits then determine the function that thelogic circuit1505 performs on its input data. Thecore logic circuit1505 then outputs the result of this function on theoutput terminal set1545.
• Any number of known logic circuits (also called logic blocks) can be used in conjunction with the invention. Examples of such known logic circuits include look-up tables (LUT's), universal logic modules (ULM's), sub-ULM's, multiplexers, and PAL's/PLA's. In addition, logic circuits can be complex logic circuits formed by multiple logic and interconnect circuits. Examples of simple and complex logic circuits can be found in Architecture and CAD for Deep-Submicron FPGAs, Betz, et al., ISBN 0792384601, 1999; and in Design of Interconnection Networks for Programmable Logic, Lemieux, et al., ISBN 1-4020-7700-9, 2003. Other examples of reconfigurable logic circuits are provided in U.S. patent application Ser. No. 10/882,583, entitled “Configurable Circuits, IC's, and Systems,” filed on Jun. 30, 2004. This Application is incorporated in the present application by reference.
• FIG. 16 illustrates an example of areconfigurable interconnect circuit1600. This interconnect circuit includes acore interconnect circuit1605 that connectsinput data terminals1610 to an output data terminal set1615 based on aconfiguration data set1620 that it receives from aswitching circuit1625, which in this example is formed by two four to onemultiplexers1640. Theswitching circuit1625 receives a larger set ofconfiguration data bits1630 that, in some embodiments, are stored in a set of storage elements1635 (e.g., a set of memory cells, such as SRAM cells). This switching circuit is controlled by a two-bit reconfiguration signal φ through twoselect lines1655. Whenever the reconfiguration signal changes, the switching circuit supplies a different set of two configuration data bits to thecore interconnect circuit1605. The configuration data bits then determine the connection scheme that theinterconnect circuit1605 uses to connect the input andoutput terminals1610 and1615.
• Any number of known interconnect circuits (also called interconnects or programmable interconnects) can be used in conjunction with the invention. Examples of such interconnect circuits include switch boxes, connection boxes, switching or routing matrices, full- or partial-cross bars, etc. Such interconnects can be implemented by using a variety of known techniques and structures. Examples of interconnect circuits can be found in Architecture and CAD for Deep-Submicron FPGAs, Betz, et al., ISBN 0792384601, 1999, and in Design of Interconnection Networks for Programmable Logic, Lemieux, et al., ISBN 1-4020-7700-9, 2003. Other examples of reconfigurable interconnect circuits are provided in the U.S. patent application Ser. No. 10/882,583.
• As mentioned above, the logic andinterconnect circuits1500 and1600 each receive a reconfiguration signal φ. In some embodiments, this signal is a sub-cycle signal that allows thecircuits1500 and1600 to reconfigure on a sub-cycle basis; i.e., to reconfigure one or more times within a cycle of a primary clock. The primary clock might be a design clock that is specified by a design (e.g., it is specified by the design in the RTL or a hardware description language (HDL)), or an interface clock that defines an i/o rate.
• Several novel techniques for distributing reconfiguration clocking signals (p are described in U.S. Patent Application entitled “Configurable IC with Interconnect Circuits that also Perform Storage Operations”, which is filed concurrently with the present application, with attorney docket number TBUL.P0022. This application is incorporated herein by reference. In conjunction with these clock distribution techniques, this application discloses several novel circuits for supplying configuration data to configurable circuits on a sub-cycle basis, based on the distributed clock signals.
• II. Configurable IC Architecture with Long Offset Direct Connections
• FIGS. 17-22 illustrate one example of the invention's architecture for a configurable or reconfigurable IC. As shown inFIG. 17, this architecture is formed by numerous rectangularconfigurable tiles1705 that are arranged in an array with multiple rows and columns. One of ordinary skill will realize that in other embodiments the tiles can have different shapes and can arranged the configurable tiles in other arrangements (e.g., the tiles might not have rectangular shapes in some embodiments).
• InFIGS. 17-22, each configurable tile includes a three-input logic circuit1710, three input-select interconnects1715,1720, and1725, and tworouting interconnects1730 and1735. As further described below, other configurable tiles can include other types of circuits, such as memory arrays instead of logic circuits.
• In thearrangement1700 ofFIG. 17, thelogic circuit1710 in each tile is a LUT, and the interconnect circuits are multiplexers. Other embodiments, however, might use other logic and/or interconnect circuits instead of or in conjunction with the LUT's and multiplexers. An input-select interconnect in some embodiments is an interconnect that has a fan out of one (i.e., its output is only provided to one circuit). In thearrangement1700, a particular tile's input-select multiplexer (IMUX) is a multiplexer that supplies one input signal of the three-input LUT1710 in the particular tile. In other words, in thearrangement1700, an input select multiplexer receives several input signals for theLUT1710 in its tile, and passes one of these input signals to its LUT.
• A routing multiplexer (RMUX) in thearrangement1700 is an interconnect circuit that can receive signals from and supply signals to interconnect and logic circuits in other tiles in the arrangement. Unlike an input select multiplexer that only provides its output to a single logic circuit (i.e., that only has a fan out of one), a routing multiplexer in some embodiments either provides its output to several logic and/or interconnect circuits (i.e., has a fan out greater than one), or provides its output to other interconnect circuits.
• Thearrangement1700 ofFIG. 17 includes numerous long offset direct connections that allow an input-select or routing multiplexer in a particular tile to receive directly signals from a routing multiplexer or a logic circuit of another tile that (1) is not a neighbor of the particular tile, and (2) is not in the same row or column in thearrangement1700 as the particular tile. Each such direct connection provides the output of a routing multiplexer or logic circuit in a first particular tile to a multiplexer (IMUX or RMUX) of a second particular tile that is separated from the first particular tile in the array either (1) by more than one row and at least one column, or (2) by more than one column and at least one row.
• For thearrangement1700,FIGS. 18-22 illustrate one example of a direct connection scheme with numerous such direct long offset direct connections. This direct connection scheme is shown for connecting the multiplexers of one tile with the LUT's and multiplexers of other tiles. This same direct connection scheme can be used for all tiles in the array, with the exception the certain provisions need to be made for tiles on or close to the boundary of the array.
• FIG. 18 illustrates that the first inputselect multiplexer1715 connects to four neighboring LUT's1805,1810,1815, and1820, two offset LUT's1825 and1830, and two offsetrouting multiplexers1835 and1840.FIG. 19 illustrates that the second inputselect multiplexer1720 connects to four neighboring offset LUT's1905,1910,1915, and1920, two other offset LUT's1925 and1930, and two offsetrouting multiplexers1935 and1940.
• FIG. 20 illustrates that the third inputselect multiplexer1725 connects to eight neighboring offset LUT's2005-2040.FIG. 21 illustrates that thefirst routing multiplexer1730 connects to the four neighboring LUT's1905,1910,1915, and1920 and to four horizontally or vertically alignedrouting multiplexers2105,2110,2115, and2120.FIG. 22 illustrates that thesecond routing multiplexer1735 connects to the four LUT's2205,2210,2215, and2220 and to four horizontally or vertically alignedrouting multiplexers2225,2230,2235, and2240.
• In the architecture illustrated inFIGS. 17-22, each tile includes one three-input LUT, three input-select multiplexers, and two routing multiplexers. Other embodiments, however, might have a different number of LUT's in each tile, a different number of inputs for each LUT, a different number of input-select multiplexers, and/or a different number of routing multiplexers.
• For instance, some embodiments might employ an architecture that has in each tile: one three-input LUT, three input-select multiplexers, and eight routing multiplexers. Table 1 below specifies one such architecture for a configurable or reconfigurable IC. Table 1 specifies the architecture by listing the inputs of the multiplexers in a particular tile and providing the source of the inputs.

  TABLE 1
  	Identity	Input	Source of Input
  Type of	of the	of the	(In terms of position
  Multiplexer	Multiplexer	Multiplexer	of corresponding tile)
  Routing	0	0	Routing multiplexer 3 of the
  			tile at position 0, −2
  			with respect to current tile
  Routing	0	1	The LUT of the tile at
  			position −2, 1 with
  			respect to current tile
  Routing	0	2	Routing multiplexer 3 of
  			the current tile
  Routing	1	0	The LUT of the tile at
  			position 4, −3 with
  			respect to current tile
  Routing	1	1	The LUT of the tile at
  			position −2, −2 with
  			respect to current tile
  Routing	1	2	The LUT of the tile at
  			position 2, 2 with
  			respect to current tile
  Routing	2	0	The LUT of the tile at
  			position 4, 0 with
  			respect to current tile
  Routing	2	1	The LUT of the tile at
  			position 0, 1 with
  			respect to current tile
  Routing	2	2	Routing multiplexer 1 of
  			the tile at position 0, 1
  			with respect to current tile
  Routing	2	3	The LUT of the current tile
  Routing	2	4	Routing multiplexer 1
  			of the current tile
  Routing	2	5	The LUT of the tile at
  			position 0, −1 with
  			respect to current tile
  Routing	2	6	Routing multiplexer 2 of
  			the tile at position −1, 0
  			with respect to current tile
  Routing	3	0	The LUT of the tile at
  			position 0, 3 with
  			respect to current tile
  Routing	3	1	The LUT of the tile at
  			position 0, 4 with
  			respect to current tile
  Routing	3	2	The LUT of the tile at
  			position 2, 0 with
  			respect to current tile
  Routing	3	3	Routing multiplexer 2 of
  			the tile at position −1, 0
  			with respect to current tile
  Routing	3	4	The LUT of the tile at
  			position 4, 4 with
  			respect to current tile
  Routing	3	5	Routing multiplexer 3 of
  			the tile at position 0, −2
  			with respect to current tile
  Routing	3	6	The LUT of the tile at
  			position 0, −2 with
  			respect to current tile
  Routing	3	7	Routing multiplexer 5
  			of the current tile
  Routing	4	0	Routing multiplexer 3
  			of the tile at position 0, 2
  			with respect to current tile
  Routing	4	1	The LUT of the current tile
  Routing	4	2	Routing multiplexer 6 of
  			the tile at position −1, 0
  			with respect to current tile
  Routing	4	3	Routing multiplexer 0
  			of the current tile
  Routing	4	4	Routing multiplexer 7 of
  			the tile at position 0, 1
  			with respect to current tile
  Routing	5	0	Routing multiplexer 3 of
  			the tile at position 0, −4
  			with respect to current tile
  Routing	5	1	The LUT of the tile at
  			position −2, 0 with
  			respect to current tile
  Routing	5	2	Routing multiplexer 2 of
  			the current tile
  Routing	6	0	The LUT of the tile at
  			position 2, 0 with
  			respect to current tile
  Routing	6	1	Routing multiplexer 1 of
  			the tile at position 0, 1
  			with respect to current tile
  Routing	7	0	Routing multiplexer 2 of
  			the tile at position −2, 0
  			with respect to current tile
  Routing	7	1	Routing multiplexer 2 of
  			the tile at position −1, 0
  			with respect to current tile
  Input-Select	0	0	Routing multiplexer 5 of
  			the tile at position 0, 1
  			with respect to current tile
  Input-Select	0	1	Routing multiplexer 5 of
  			the current tile
  Input-Select	0	2	Routing multiplexer 2 of
  			the tile at position −2, 0
  			with respect to current tile
  Input-Select	0	3	Routing multiplexer 4 of
  			the tile at position −8, 0
  			with respect to current tile
  Input-Select	0	4	Routing multiplexer 4 of
  			the tile at position 5, 3
  			with respect to current tile
  Input-Select	0	5	Routing multiplexer 4 of
  			the tile at position −7, 0
  			with respect to current tile
  Input-Select	0	6	Routing multiplexer 4 of
  			the tile at position 8, 0
  			with respect to current tile
  Input-Select	0	7	Routing multiplexer 4 of
  			the tile at position 2, 0
  			with respect to current tile
  Input-Select	1	0	Routing multiplexer 4 of
  			the tile at position 0, 2
  			with respect to current tile
  Input-Select	1	1	Routing multiplexer 4 of
  			the tile at position −4, 0
  			with respect to current tile
  Input-Select	1	2	Routing multiplexer 3 of
  			the tile at position 0, −4
  			with respect to current tile
  Input-Select	1	3	Routing multiplexer 4 of
  			the tile at position −4, 3
  			with respect to current tile
  Input-Select	1	4	Routing multiplexer 2 of
  			the current tile
  Input-Select	1	5	Routing multiplexer 4 of
  			the tile at position 7, 0
  			with respect to current tile
  Input-Select	1	6	Routing multiplexer 4 of
  			the tile at position 7, −1
  			with respect to current tile
  Input-Select	1	7	Routing multiplexer 4 of
  			the tile at position 4, 4 with
  			respect to current tile
  Input-Select	2	0	Routing multiplexer 0 of
  			the current tile
  Input-Select	2	1	LUT of the tile at
  			position −2, 0 with
  			respect to current tile
  Input-Select	2	2	LUT of the tile at
  			position 2, −2 with
  			respect to current tile
  Input-Select	2	3	Routing multiplexer 2 of
  			the tile at position −2, 0
  			with respect to current tile
  Input-Select	2	4	Routing multiplexer 5 of
  			the tile at position 0, 1
  			with respect to current tile
  Input-Select	2	5	Routing multiplexer 6 of
  			the current tile
  Input-Select	2	6	Routing multiplexer 4 of
  			the tile at position −2, 0
  			with respect to current tile
  Input-Select	2	7	LUT of the tile at position
  			4, −2 with respect to current tile

• As mentioned above, Table 1 specifies the architecture by listing the inputs of the multiplexers in a particular tile and providing the source of the inputs. The source of each input is expressed as (1) a component in the particular tile, or (2) a component in another tile, which is identified in terms of two coordinates (a,b) that express the location of the other tile by reference to the location of the particular tile. These two coordinates are defined in a coordinate system that has the particular tile as its origin. In this coordinate system, each unit along its x- or y-axis is one tile. For instance, using this notation, thetile1850 inFIG. 18 is connected to the following tiles: (1)tile1855 atlocation 1,0, (2)tile1860 atlocation 0,1, (3)tile1865 atlocation tile1870 atlocation 0,−1, (5)tile1875 atlocation 2,2, and (6)tile1880 at location −2,−2.
• Table 2 specifies another embodiment's architecture for a configurable or reconfigurable IC. In this embodiment, each tile has one three-input LUT, three input-select multiplexers, and six routing multiplexers. Table 2 specifies the IC architecture by using the same nomenclature as Table 1.

  TABLE 2
  	Identity	Input	Source of Input
  Type of	of the	of the	(In terms of position
  Multiplexer	Multiplexer	Multiplexer	of corresponding tile)
  Routing	0	0	The LUT of the tile at
  			position 2, −1 with
  			respect to current tile
  Routing	0	1	The LUT of the tile at
  			position −4, 3 with
  			respect to current tile
  Routing	0	2	Routing multiplexer 1 of
  			the tile at position −4, 0
  			with respect to current tile
  Routing	0	3	Routing multiplexer 5 of
  			the tile at position 0, −1
  			with respect to current tile
  Routing	0	4	Routing multiplexer 0 of
  			the tile at position 7, 0
  			with respect to current tile
  Routing	0	5	Routing multiplexer 4 of
  			the tile at position 0, −2
  			with respect to current tile
  Routing	0	6	Routing multiplexer 0 of
  			the tile at position −4, 0
  			with respect to current tile
  Routing	0	7	Routing multiplexer 3 of
  			the tile at position −3, 0
  			with respect to current tile
  Routing	1	0	The LUT of the tile at
  			position −2, −1 with
  			respect to current tile
  Routing	1	1	The LUT of the tile at
  			position −5, 3 with
  			respect to current tile
  Routing	1	2	The LUT of the tile at
  			position 5, −2 with
  			respect to current tile
  Routing	1	3	Routing multiplexer 1 of
  			the tile at position 0, 3
  			with respect to current tile
  Routing	1	4	Routing multiplexer 3 of
  			the tile at position 0, −1
  			with respect to current tile
  Routing	1	5	Routing multiplexer 3 of
  			the tile at position −5, 3
  			with respect to current tile
  Routing	1	6	Routing multiplexer 4 of
  			the tile at position 0, 1
  			with respect to current tile
  Routing	1	7	Routing multiplexer 4 of
  			the tile at position 0, −2
  			with respect to current tile
  Routing	2	0	The LUT of the tile at
  			position −1, −1 with
  			respect to current tile
  Routing	2	1	The LUT of the tile at
  			position −1, 3 with
  			respect to current tile
  Routing	2	2	Routing multiplexer 2 of
  			the tile at position −1, 0
  			with respect to current tile
  Routing	2	3	Routing multiplexer 3 of
  			the tile at position −3, 2
  			with respect to current tile
  Routing	2	4	Routing multiplexer 0 of
  			the tile at position −1, 1
  			with respect to current tile
  Routing	2	5	Routing multiplexer 4 of
  			the tile at position −8, 0
  			with respect to current tile
  Routing	2	6	Routing multiplexer 2 of
  			the tile at position 0, −1
  			with respect to current tile
  Routing	2	7	The LUT of the tile at
  			position 5, −2 with
  			respect to current tile
  Routing	3	0	The LUT of the tile at
  			position −2, −1 with
  			respect to current tile
  Routing	3	1	The LUT of the tile at
  			position 1, 3 with
  			respect to current tile
  Routing	3	2	The LUT of the tile at
  			position −3, −2 with
  			respect to current tile
  Routing	3	3	Routing multiplexer 1 of
  			the tile at position −2, 0
  			with respect to current tile
  Routing	3	4	Routing multiplexer 0 of
  			the current tile
  Routing	3	5	Routing multiplexer 1 of
  			the tile at position 6, −1
  			with respect to current tile
  Routing	3	6	Routing multiplexer 4 of
  			the tile at position 0, −1
  			with respect to current tile
  Routing	3	7	Routing multiplexer 0 of
  			the tile at position 1, −5
  			with respect to current tile
  Routing	4	0	Routing multiplexer 4 of
  			the tile at position −4, 0
  			with respect to current tile
  Routing	4	1	Routing multiplexer 4 of
  			the tile at position 4, 0
  			with respect to current tile
  Routing	4	2	Routing multiplexer 3 of
  			the tile at position −2, 0
  			with respect to current tile
  Routing	4	3	Routing multiplexer 3 of
  			the tile at position −1, −3
  			with respect to current tile
  Routing	4	4	Routing multiplexer 0 of
  			the tile at position 7, 0
  			with respect to current tile
  Routing	4	5	Routing multiplexer 3 of
  			the tile at position −6, −1
  			with respect to current tile
  Routing	4	6	Routing multiplexer 5 of
  			the tile at position 4, 2
  			with respect to current tile
  Routing	4	7	The LUT of the tile at
  			position 0, 2 with
  			respect to current tile
  Routing	5	0	Constant Input
  Routing	5	1	Constant Input
  Routing	5	2	Routing multiplexer 4 of
  			the tile at position 1, 0
  			with respect to current tile
  Routing	5	3	Routing multiplexer 3 of
  			the tile at position 6, 2
  			with respect to current tile
  Routing	5	4	Routing multiplexer 1 of
  			the tile at position −4, 0
  			with respect to current tile
  Routing	5	5	Routing multiplexer 1 of
  			the tile at position −1, −1
  			with respect to current tile
  Routing	5	6	Routing multiplexer 0 of
  			the tile at position 1, 0
  			with respect to current tile
  Routing	5	7	Routing multiplexer 0 of
  			the tile at position 7, 0
  			with respect to current tile
  Input-Select	0	0	Routing multiplexer 4
  			of current tile
  Input-Select	0	1	Routing multiplexer 4
  			of the current tile
  Input-Select	0	2	Routing multiplexer 1 of
  			the tile at position 0, 1
  			with respect to current tile
  Input-Select	0	3	Routing multiplexer 5 of
  			the tile at position 1, 1
  			with respect to current tile
  Input-Select	0	4	Routing multiplexer 5 of
  			the tile at position 0, −5
  			with respect to current tile
  Input-Select	0	5	Routing multiplexer 3 of
  			the tile at position 0, 2
  			with respect to current tile
  Input-Select	0	6	Routing multiplexer 1 of
  			the tile at position −3, 0
  			with respect to current tile
  Input-Select	0	7	The LUT of the tile at
  			position 0, −1 with
  			respect to current tile
  Input-Select	1	0	Routing multiplexer 0 of
  			the tile at position 4, 0
  			with respect to current tile
  Input-Select	1	1	Routing multiplexer 1 of
  			the tile at position 4, 0
  			with respect to current tile
  Input-Select	1	2	The LUT of the tile at
  			position −2, −2
  			with respect to current tile
  Input-Select	1	3	Routing multiplexer 5 of
  			the tile at position 0, −3
  			with respect to current tile
  Input-Select	1	4	Routing multiplexer 4 of
  			the tile at position 0, −1
  			with respect to current tile
  Input-Select	1	5	Routing multiplexer 4 of
  			the tile at position 1, 0
  			with respect to current tile
  Input-Select	1	6	Routing multiplexer 4
  			of the current tile
  Input-Select	1	7	Routing multiplexer 1 of
  			the tile at position −1, 5
  			with respect to current tile
  Input-Select	2	0	Routing multiplexer 2 of
  			the tile at position −1, 0
  			with respect to current tile
  Input-Select	2	1	Routing multiplexer 3 of
  			the tile at position −4, 0
  			with respect to current tile
  Input-Select	2	2	Routing multiplexer 0 of
  			the tile at position −1, 3
  			with respect to current tile
  Input-Select	2	3	Routing multiplexer 1 of
  			the tile at position −1, 9
  			with respect to current tile
  Input-Select	2	4	Routing multiplexer 3 of
  			the tile at position 0, −7
  			with respect to current tile
  Input-Select	2	5	Routing multiplexer 0 of
  			the tile at position 0, −4
  			with respect to current tile
  Input-Select	2	6	The LUT of the tile at
  			position 1, −1 with
  			respect to current tile
  Input-Select	2	7	The LUT of the tile at
  			position −1, 2 with
  			respect to current tile

• In some embodiments, each particular tile has the same exact direct connections listed above in Table 1 or 2, with the exception perhaps of tiles at or close to the boundary of the tile arrangement. In some embodiments, the tiles at or close to the boundary do not have some of the direct connections that extend past the boundary. Some embodiments “stitch” together tiles that are at or close to the tile array boundary, by defining unique direct connections between such tiles, where these unique direct connections take the place of the direct connections that would otherwise extend past the tile array boundary.
• In other embodiments, the tiles at or close to the boundary do have the same direct connection but these direct connections wrap around to the other side of the tile arrangement. For instance, when a tile is on the top of the tile array and it has a routing multiplexer that is suppose to connect to a tile above it, the direct connection might be eliminated or it might be made with a tile at the bottom of the tile array.
• In some embodiments, the direct connections illustrated inFIGS. 17-22, and in Table 1 or 2, are the direct connections of each computational tile (with the possible exception of computational tiles at the boundary), but not the direct connection of the non-computational tiles (e.g., a tile that includes a memory). In other embodiments, the direct connections illustrated in Table 1 or 2 are the direct connections of some or all computational and non-computational tiles.
• The architecture of some embodiments includes one or more loops between the output of a LUT in a particular computational tile and its input. For instance, the architecture defined by Table 2 includes three such loops, one for each input of the 3-input LUT. Each such loop is established through two routing multiplexers of two other tiles and the input select multiplexer of the LUT. In this manner, the output of the LUT can be stored in a user register formed by routing multiplexers that can be enabled to serve as latches, and this output can be fedback to the LUT's input.
• Routing multiplexer 5 in the architecture specified by Table 2 receives two constant values (e.g., receives a “0” and a “1”). This routing multiplexer has connections withrouting multiplexers 1, 3, and 4. Theserouting multiplexers 1, 3, and 4 have good connectivity with the input select multiplexers. As further mentioned below in Section IV, the input select multiplexers are hybrid logic/interconnect circuits in some embodiments. Some embodiments use these hybrid structures to decompose and implement logic functions, as described in U.S. Patent Application entitled “Hybrid Configurable Circuit for Configurable IC”, filed concurrently with the present application, with attorney docket number TBUL.P0010. As described in this application, these hybrid structures need to receive constant values in some instances when they are decomposing and implementing logic functions. Hence, the architecture illustrated in Table 2 feeds constant values to each routing multiplexer 5 of some or all computational tiles. These constant values can then be selectively routed to input-select hybrid multiplexers (through the multiplexers5, andmultiplexers 1, 3, and 4), which then use them during their decompose and implement logic functions.
• In some embodiments, the LUT's, IMUX's, and RMUX's in all the tiles are configurable circuits. Also, in some embodiments, all these circuits are sub-cycle configurable circuits that receive their configuration data on a sub-cycle basis. For instance, each sub-cycle configurable LUT or multiplexer receives its configuration data on a sub-cycle basis through a novel two-tier multiplexer structure described in the above-mentioned U.S. Patent Application entitled “Configurable IC with Interconnect Circuits that also Perform Storage Operations”, which is filed concurrently with the present application, with attorney docket number TBUL.P0022.
• In other embodiments, not all the LUT's, IMUX's, and RMUX's of a configurable IC are configurable or sub-cycle reconfigurable. For instance, in some embodiments, only the IMUX's and RMUX's are configurable or sub-cycle reconfigurable, while the LUT's are only configurable and not sub-cycle reconfigurable.
• Also, tiles were described above to include LUT's, IMUX's, and RMUX's. In some embodiments, tiles also include other circuits as further described below. Also, as further described in the above-incorporated U.S. Patent Application entitled “Configurable IC with Interconnect Circuits that also Perform Storage Operations” (which is filed concurrently with the present application, with attorney docket number TBUL.P0022) these tiles include local sub-cycle signal generators in some embodiments. Such sub-cycle signal generators generate sub-cycle signals for retrieving configuration data sets from memory storage. In some embodiments, these generators generate their sub-cycle signals based on globally distributed clock signals.
• Tiles can also include memory arrays in conjunction with the LUT's, IMUX's, and RMUX's, or instead of some of these circuits (e.g., the LUT's). Several such tiles will be further described below.
• III. Asymmetric Architecture
• Some embodiments provide an asymmetric architecture for a configurable IC. In a tile-based architecture that includes routing interconnects, input-select interconnects, and logic circuits, the architecture can be asymmetric when it comes to the inputs of the routing interconnects, the outputs of the routing interconnects, the inputs of the input-select interconnects, or the output of the logic circuits. The architecture of the configurable IC of some embodiments is asymmetric with respect to all these conditions, while the architecture of other embodiments is asymmetric with respect to only some of these conditions.
• For instance, an architecture can be asymmetric with respect to the inputs of the routing interconnects when at least one input of the routing interconnect in a particular tile is not “symmetric” with any other input of the routing interconnects of the particular tile. Two inputs are symmetric when they originate from two tiles that have a symmetric relationship with respect to each other when viewed from the position of the particular tile. Some embodiments define two tiles as having a symmetric relationship with respect to the position of a third tile when the two tiles can swap positions when they are flipped about an origin that is defined at the position of the third tile. Instead of, or in conjunction with, this definition, some embodiments define two tiles as having a symmetric relationship when one tile can take the position of the other tile if the two tiles are rotated about the origin that is defined at the position of the third tile.
• FIG. 23 illustrates an example of an architecture2300 that is asymmetric with respect to the inputs of the routing interconnects. This architecture is similar to the architecture illustrated inFIGS. 17-22, except that it includes two routing-interconnect inputs2305 and2310 that are not symmetric with any of the other inputs to therouting interconnect1730. Theinput2305 comes from a routing multiplexer intile2315 at (2,3), while theinput2310 comes from a routing multiplexer intile2320 at (−1,−2). These two inputs take the place of the inputs illustrated inFIG. 21 from therouting multiplexers2115 and2120.
• Similarly, an architecture can be asymmetric with respect to the outputs of the routing interconnects of a tile when at least one output of the routing interconnect in a particular tile is not “symmetric” with any other output of the routing interconnects of the particular tile. Two outputs of one or two routing interconnects in a particular tile are asymmetric when they are supplied to two circuits at two locations in the tile arrangement that do not have a symmetric relationship with respect to each other in the configurable IC when viewed from the position of the particular tile.
• An architecture can also be asymmetric with respect to the inputs of the input-select interconnects when at least one input of the input-select interconnect in a particular tile is not “symmetric” with any other input of the input-select interconnects of the particular tile. Two inputs of one or two input-select interconnects in a particular tile are asymmetric when they are received from two circuits at two locations in the tile arrangement that do not have a symmetric relationship with respect to each other in the configurable IC when viewed from the position of the particular tile.
• An architecture can also be asymmetric with respect to the outputs of the set of logic circuits of a tile when at least one output of a logic circuit in a particular tile is not “symmetric” with any other output of the logic circuit set of the particular tile. Two outputs of one or two logic circuits in a particular tile are asymmetric when they are supplied to two circuits at two locations in the tile arrangement that do not have a symmetric relationship with respect to each other in the configurable IC when viewed from the position of the particular tile.
• As mentioned above, each tile in some embodiments has the same set of asymmetric connections (e.g., asymmetric inputs to RMUX's, asymmetric inputs to IMUX's, etc.) with other tiles, except for tiles that are at or close to the boundary of the tile arrangement that need to address boundary conditions. In other embodiments, different tiles have different sets of connections with other tiles. However, in some of these embodiments, large sets of tiles (e.g., hundreds, thousands, etc.) have the same set of asymmetric connections with other tiles. The tiles in such large sets might all be interior tiles, or they might be tiles at or close to the boundary that need to have special connections defined to address boundary issues as mentioned above. By avoiding symmetric sets of direct connections, or using only a few of them, some embodiments reduce the number of redundant cyclic direct connections in a design. Moreover, the use of direct asymmetric offset connections in these architectures increases the interconnectivity between the circuits of the IC.
• In some embodiments, the outputs or inputs of a particular tile's routing interconnects, input-select interconnects, or logic circuits are not physically symmetric as they include at least one output or one input that is not symmetric with respect to any of the other outputs or inputs. However, in some of these embodiments, the outputs or inputs of the particular tile routing interconnects, input-select interconnects, or logic circuits are isotropic or approximately isotropic. Each output or input connection can be represented in terms of a vector that is defined in terms of the start and end points of the connection. For instance, an output connection from a first routing interconnect in a first tile might take the output of the first routing interconnect to an input of a second routing interconnect in a second tile that is two tiles above and three tiles to the right of the first tile. This connection can be represented by a vector representation (3,2). A set of outputs or inputs connections is isotropic when the sum of the vectors that these connections represent equals a vector (0,0).
• IV. Routing and Input Multiplexers as Interconnect/Storage Circuits and as Hybrid Interconnect/Logic Circuits
• A. Interconnect/Storage Circuits
• Numerous of the above-described architectures use routing multiplexers. In some embodiments, some or all of these routing multiplexers are interconnect/storage circuits that are useful for maintaining state information in a configurable IC. To illustrate the need for such state elements,FIGS. 24-27 present an example of implementing an IC design with a sub-cycle reconfigurable IC.
• FIG. 24 illustrates a set of Boolean gates that compute two functions based on a set of inputs A0, B0, A1, B1, A2, and B2. The set of Boolean gates has to compute these two functions based on the received input set in one design cycle. In this example, one design cycle lasts 10 ns, as the design clock's frequency is 100 MHz. However, in this example, each gate can operate at 400 MHz. Hence, each design cycle can be broken down into four sub-cycles of 2.5 ns duration, in order to allow meet the design clock frequency of 100 MHz.
• FIG. 25 illustrates thedesign2400 ofFIG. 24 after its gates have been placed into four groups. These gates have been placed into four groups in order to break down thedesign2400 into four separate groups of gates that can be configured and executed in four sub-cycles by a smaller group of gates. The groupings illustrated inFIG. 25 are designed to separate out the computation of different sets of gates while respecting the operational dependencies of other gates. For instance,gates2405,2410, and2415 are defined as a separate group fromgates2420,2425, and2430, as these two sets of gates have no operational dependencies (i.e., the output of the gates in one set is not dependent on the output of the gates in the other set). As these two sets of gates have no operational dependencies, one set is selected for computation during the first sub-cycle (i.e., during phase 1), while the other set is selected for computation during the second sub-cycle (i.e., during phase 2). On the other hand,gates2435,2440, and2445 are dependent on the outputs of the first two sets of gates. Hence, they are designated for configuration and execution during the third sub-cycle (i.e., during phase 3). Finally, thegate2450 is dependent on the output of the first and third sets of gates, and thus it is designated for configuration and execution during the fourth sub-cycle (i.e., during phase 4).
• FIG. 26 illustrates another representation of thedesign2400 ofFIG. 24. LikeFIG. 25, the schematic inFIG. 26 illustrates four phases of operation. However, now, each gate in thedesign2400 has been replaced by a sub-cycleconfigurable logic circuit2605,2610, or2615. Also, only threelogic circuits2605,2610, and2615 are used inFIG. 26, as each of the gates inFIG. 24 can be implemented by one logic circuit, and the groupings illustrated inFIGS. 25 and 26 require at most three gates to execute during any given phase. (InFIG. 26, each logic circuit's operation during a particular phase is identified by a superscript; so, for example,reference numbers2605¹,2605², and2605³, respectively, identify the operation of thelogic circuit2605 duringphases 1, 2, and 3.)
• As shown inFIG. 26, the outputs of certain logic circuits in earlier phases need to be supplied to logic circuit operations in the later phases. Such earlier outputs can be preserved for later computations by using state elements (such as registers or latches). Such state elements (not shown) can be standalone circuits or can be part of one or more interconnect circuits.
• As mentioned above, the state elements in some embodiments are routing multiplexers that can serve as both storage and interconnect circuits. Specifically, each such routing multiplexer is a configurable interconnect/storage circuit that can be configured to act as an interconnect circuit or as a storage circuit. In some embodiments, all the routing multiplexers of a configurable or reconfigurable IC are configurable interconnect/storage circuits, while in other embodiments only some of the routing multiplexers of the IC are configurable interconnect/storage circuits.
• FIG. 27 illustrates a circuit representation of an interconnect/storage circuit2700 that can be used to implement the routing multiplexer of some embodiments. Thiscircuit2700 is formed by placing alatch2705 at the output stage of amultiplexer2710. Thelatch2705 receives a latch enable signal. When the latch enable signal is inactive, the circuit simply acts as an interconnect circuit. On the other hand, when the latch enable signal is active, the circuit acts as a latch that outputs the value that the circuit was previously outputting while serving as an interconnect circuit. Accordingly, when a second circuit in a second later configuration cycle needs to receive the value of a first circuit in a first earlier configuration cycle, thecircuit2700 can be used to receive the value in a cycle before the second later configuration cycle (e.g., in the first earlier cycle) and to latch and output the value to the second circuit in the second later sub-cycle. Thecircuit2700 and other interconnect/storage circuits are further described in the above-mentioned U.S. Patent Application entitled “Configurable IC with Interconnect Circuits that also Perform Storage Operations”, which is filed concurrently with the present application, with attorney docket number TBUL.P0022.
• Some embodiments do not use the interconnect/storage circuits (such as thecircuit2700 ofFIG. 27) for any of the input-select multiplexers. Other embodiments, however, use such interconnect/storage circuits for some or all of the input-select multiplexers. Yet other embodiments might use the interconnect/storage circuits for only the input-select multiplexers, and not for the routing multiplexers.
• B. Hybrid Circuits
• The configurable IC's of some embodiments include numerous input select multiplexers that are hybrid multiplexers, called HUMUX's. An HUMUX is a multiplexer that can receive “user-design signals”, configuration data, or both user-design signals and configuration data for its select signals. A user-design signal within a configurable IC is a signal that is generated by a circuit (e.g., a logic circuit) of the configurable IC. The word “user” in the term “user-design signal” connotes that the signal is a signal that the configurable IC generates for a particular user application. User-design signal is abbreviated to user signal in some of the discussion below.
• In some embodiments, a user signal is not a configuration or clock signal that is generated by or supplied to the configurable IC. In some embodiments, a user signal is a signal that is a function of at least a portion of the configuration data received by the configurable IC and at least a portion of the inputs to the configurable IC. In these embodiments, the user signal can also be dependent on (i.e., can also be a function of) the state of the configurable IC. The initial state of a configurable IC is a function of the configuration data received by the configurable IC and the inputs to the configurable IC. Subsequent states of the configurable IC are functions of the configuration data received by the configurable IC, the inputs to the configurable IC, and the prior states of the configurable IC.
• FIG. 28 illustrates anHUMUX2800. This HUMUX includes two two-to-one multiplexers2820, a four-to-one multiplexer2825, a set ofinput terminals2805, anoutput terminal2810, and a set ofselect terminals2815. From the outside, the HUMUX looks like a four-to-one multiplexer that has fourdata inputs2805, onedata output2810, and fourselect terminals2815. Also, from the outside, the HUMUX looks like it passes one of its fourdata inputs2805 to its onedata output2810 based on the value of two of the four signals that it receives along its fourselect lines2815.
• Internally, the two two-to-one multiplexers2820 pass two of the signals from the fourselect lines2815 to the twoselect terminals2840 of the four-to-one multiplexer2825. As shown inFIG. 28, each two-to-one multiplexer2820 receives two input signals, which include one user-design signal and one stored configuration signal stored in astorage element2845. Each of the two-to-one multiplexers2820 outputs one of the two input signals that it receives based on the configuration bit that it receives along its select line2850.
• AlthoughFIG. 28 illustrates two configuration bits stored in two storage elements, some embodiments drive both multiplexers2820 off one configuration bit that is stored in one storage element. Also, some embodiments have a sub-set of theselect lines2840 always driven by configuration data. In other words, these embodiments drive only one of theselect lines2840 potentially with a user signal; the otherselect line2840 would always be driven by configuration data. These and other HUMUX structures are described in U.S. Patent Application entitled “Hybrid Configurable Circuit for Configurable IC”, filed concurrently with the present application, with attorney docket number TBUL.P0010. This application is incorporated herein by reference.
• The two signals output by the two multiplexers2820 then serve as the select signals of themultiplexer2825, and thereby direct thismultiplexer2825 to output online2810 one of the four input signals that it receives onlines2805. The two multiplexers2820 can output onlines2840 either two user-design signals, two configuration signals, or one user-design signal and one configuration signal. Accordingly, through the two multiplexers2820, the operation of themultiplexer2825 can be controlled by two user-design signals, two configuration signals, or a mix of user/configuration signals.
• HUMUX's are hybrid interconnect/logic circuits. In other words, HUMUX's can serve as logic and interconnect circuits in a configurable IC. This hybrid quality is especially advantageous since, as logic circuits, HUMUX's can be used to decompose and implement functions. In order to decompose and implement functions with HUMUX's, some embodiments define one input of some or all HUMUX's to be a permanently inverting input. The use of an HUMUX to decompose functions is further described in the above-incorporated U.S. Patent Application entitled “Hybrid Configurable Circuit for Configurable IC”, filed concurrently with the present application, with attorney docket number TBUL.P0010.
• This incorporated application also further describes the use of HUMUX's for some or all of the input select multiplexers. It further describes the use of HUMUX's as some or all of the routing multiplexers. Some embodiments, however, use HUMUX's only for some or all of the input select multiplexers, while using the interconnect/storage circuit ofFIG. 27 for some or all of the routing multiplexers.
• V. Architecture with Fast Carry Chains
• In some embodiments, the examples illustrated inFIGS. 17-22 and Tables 1 and 2 define the physical architecture of a configurable IC. In other embodiments, these examples topologically illustrate the architecture of a configurable IC. Specifically, in these embodiments, the direct connections illustrated and defined inFIGS. 18-22 and Tables 1 and 2 are only meant to show direct connections between the circuits in the configurable IC, without specifying (1) a particular geometric layout for the wire segments that establish the direct connections, or even (2) a particular position of the circuits.
• In some embodiments, the position and orientation of the circuits in the actual physical architecture of a configurable IC is different from the position and orientation of the circuits in the topological architecture of the configurable IC. Accordingly, in these embodiments, the IC's physical architecture appears quite different from its topological architecture.
• FIGS. 29 and 30 provide one example that illustrates such a difference. Specifically,FIG. 29 topologically illustrates a portion of the architecture of a configurable IC2900. This IC's architecture inFIG. 29 is formed by a series of tiles that are arranged in multiple topological rows and columns. InFIG. 29, each tile is numbered. Like eachtile1705 inFIG. 17, eachtile2905 inFIG. 29 includes tworouting multiplexers1730 and1735, three input-select multiplexers1715,1720, and1725, and one threeinput LUT1710.
• However, unlikeFIG. 17,FIG. 29 also illustrates acarry logic circuit2910 in each tile. The LUT and carry logic circuit in each tile form a logic carry block (LCB) that allows the LUT to implement an adder/subtractor, which can perform an add or subtract operation as further described below.FIG. 30 illustrates a portion of the actual physical architecture of the configurable IC2900. As shown in this figure, the configurable IC2900 is formed by (1) grouping sets of four topologically adjacent tiles that are in the same topological row inFIG. 29, and (2) aligning the tiles in each group so that their logic carry blocks are adjacent to each other. In each group of aligned tiles, the tiles are rotated by −90° or 90° with respect to the alignment illustrated inFIG. 29. Each set of four aligned tiles forms an aligned tile layout that has four logic circuits and four carry logic circuits that are close to each other.
• Specifically, in this example, (1) the first topological row is divided into a first set of tiles1-4 and a second set of tiles5-8, (2) the second topological row is divided into a third set of tiles9-12 and a fourth set of tiles13-16, (3) the third topological row is divided into a fifth set of tiles17-20 and a sixth set of tiles21-24, and (4) the fourth topological row is divided into a seventh set of tiles25-28 and an eighth set of tiles29-32.In each set of four tiles, the first two tiles are rotated by −90° with respect to the alignment illustrated inFIG. 29, while the second two tiles are rotated by 90° with respect to the alignment illustrated inFIG. 29. The tiles in each set are aligned in the manner illustrated inFIG. 30, to form aligned tile layouts. For instance, tiles1-4 form a first aligned tile layout, tiles5-8 form a second aligned tile layout, and so on.
• In some embodiments, the aligned tile layout can be viewed as a layout replica (i.e., unit of architectural regularity) that is defined collectively as a set, and that is repeated across the layout of the configurable IC. In some embodiments, tile layouts can actually be used as layout replicas during the design process to define the layout of a configurable IC, while in other embodiments tile layouts are simply an abstraction for viewing a pattern of circuits that is repeated across the layout.
• Having the aligned tile layout with the same circuit elements simplifies the process for designing and fabricating the IC, as it allows the same circuit designs and mask patterns to be repetitively used to design and fabricate the IC. In some embodiments, the similar aligned tile layout not only has the same circuit elements but also have the same exact internal wiring between their circuit elements. Having such a layout further simplifies the design and fabrication processes as it further simplifies the design and mask making processes.
• To further elaborate on the proximity of the logic carry blocks within each aligned tile layout,FIG. 31 provides another illustration of an alignedtile layout3100, which is formed by four tiles3105-3120 (in a topological row) that are aligned in the physical architecture. In this illustration, only the logic carry blocks3125-3140 within each tile is illustrated. As mentioned above, each LCB is formed by a LUT and its associated carry logic circuit in a tile. As shown inFIG. 31, the alignment of the tiles clusters the logic carry blocks3125-3140 close to each other. This close proximity, in turn, allows the four LCB's to form a fast nibble wide (4-bit) adder/subtractor.
• To elaborate on this,FIG. 32 provides a simple illustration of anLCB3200 that is formed by a three-input LUT3205 and its associatedcarry logic circuit3210. When acting as a one-bit adder/subtractor, a three-input LUT3205 receives (1) two one-bit inputs “a” and “b” to add and (2) a carry signal “c” (C_IN) that gets factored in the addition. TheLCB3200 of a particular tile can receive (1) a local carry signal from the carry logic circuit of a neighboring tile in the same topological row as the particular tile, or (2) a global carry signal from a carry logic circuit in a different topological row.
• Based on the three input signals that it receives, theLUT3205 expresses the result of its addition operation in terms of a function f(a,b,c), a propagate signal P, and a generate signal G. When theLUT3205 acts as an adder/subtractor, the function f(a,b,c) expresses the sum of “a” and “b” with “c” (C_IN) as the carry-in signal. More specifically, when adding two one-bit values, the LUT's output function f(a,b,c) computes the sum as (a⊕b)⊕c. When subtracting a one-bit value , the LUT's output function f(a,b,c) computes a “2's complement” subtraction as (a⊕b)⊕c.
• Also, when theLCB3200 adds two one-bit values, the propagate signal P equals (a⊕b), and the generate signal G equals (a·b). Alternatively, when theLCB3200 subtracts two one-bit values, the propagate signal P equals (a⊕b), and the generate signal G equals (a·b). The propagate and generate signals are supplied to thecarry logic circuit3210, which, based on these signals, computes a carry signal C_OUTthat equals G+(P·c). The generate signal directs thecarry logic circuit3210 to generate a carry signal C_OUT, regardless of whether there is a carry that is being propagated. The propagate signal directs thecarry logic circuit3210 to propagate the carry signal regardless of whether there is a carry that is being generated. The carry signal C_OUTcomputed by thecircuit3210 is the next most significant LCB in a ripple chain of adders that add two multi bit values, is the most significant bit of the resulting add operation, or is the expressed overflow.
• Each LCB can form a one-bit adder/subtractor or form larger adders/subtractors when it is used in conjunction with other LCB's. Accordingly, to form fast four-bit adders/subtractors, some embodiments place the four LCB's in an aligned tile layout close to each other, so that the carry signals can be quickly passed between adjacent LCB's.FIG. 31 shows acarry signal trace3150 that highlights the direction of carry-signal flow through four, aligned LCB's of an aligned tile layout. Alternative carry-signal flows through four, aligned LCB's are also possible, such as the flow illustrated inFIG. 33. Due to the proximity of the LCB's, most of these carry-signal flows allow the four, aligned LCB's to form a fast nibble-wide adder/subtractor. In addition, when ganged with other fast nibble wide adders/subtractors that are on the same topological row, the nibble wide adders/subtractors can form fast byte-wise adders/subtractors (as shown inFIG. 34) or other larger adders/subtractors (sixteen bit adders/subtractors, thirty-two bit adders/subtractors, etc.).
• As mentioned above,FIG. 29 provides a topological illustration of a portion of a configurable IC's architecture. The description above highlighted that in some embodiments the position and orientation of the circuits in the actual physical architecture of the configurable IC is different from the position and orientation of the circuits in the topological architecture of the configurable IC. Also, in some embodiments, the topological and/or actual geometric layout of wire segments and/or vias that define the direct connections between the circuits can change once the tiles are grouped and aligned.
• To illustrate this,FIG. 29 presentstopological illustrations2915 and2920 of two direct connections, one between the second routing multiplexers oftiles1 and26, and one between the second routing multiplexers oftiles2 and27.FIG. 30 presentstopological illustrations3015 and3020 of the same two direct connections after the tiles have been grouped and aligned. As shown in these two figures, the realignment of the tiles changes the topological direct connections by changing the relative position of the two circuits that are connected in each connected pair of circuits.
• The change in the relative position of the connected circuit pairs will typically also result in a change in the actual geometric layout of the direct connection between the connected circuits. As mentioned above, the geometric layout of a direct connection often differs from its topological representation. In addition, as mentioned above, a direct connection between two circuits can be achieved by (1) a set of wire segments that traverse through a set of the wiring layers of the IC, and (2) a set of vias when two or more wiring layers are involved. A direct connection can also include one or more buffers in some embodiments, while such a connection does not include buffers in other embodiments.
• VI. Architecture with Shared Carry Logic
• Instead of having to group and align tiles, some embodiments define aligned tile layouts from the start and then simply use the notion of tiles within the aligned tile layouts to define the interconnect topology of the circuits. Some of these embodiments specify the position of the four LUT's and four carry logic circuits within each aligned tile layout to be close to each other so that these LUT's and circuits can form fast nibble wide adders/subtractors.
• Alternatively, in an aligned tile layout, some embodiments define one common carry chain that is shared among the four logic circuits in the tile layout.FIG. 35 illustrates onesuch layout3500. As shown in this figure, this layout includes four logic circuits (0-3), and a sharedcarry logic3505.
• Each logic circuit i receives three input signals a_i, b_i, c_ithrough three input-select multiplexers3550 During an add operation, the third input c_iof each LUT is one of the outputs of thecarry logic3505. Based on the three input signals that it receives, each LUT i expresses the result of its addition operation in terms of (1) a function f_i(a_i, b_i, c_i) that is dependent on the three input signals, (2) a propagate signal P_ithat equals (a_i⊕b_i) when a_iand b_iare added and equals (a_i⊕b_i) when b_iis subtracted from a_i, and (3) a generate signal G_ithat equals (a_i·b_i) when a_iand b_iare added and equals (a_i·b_i) when b_iis subtracted from a_i,
• Also, during an add or subtract operation, each LUT i provides its propagate signal P_i. and generate signal G_ito thecarry logic3505. Thecarry logic3505 also receives a carry input C_IN, which is either a local carry input C_INL(i.e., a carry input from a tile in the same topological row) or a global carry input C_ING(i.e., a carry input from a tile in a different topological row), as determined by amultiplexer3510 associated with the aligned tile group.
• Based on its input signals, thecarry logic3505 generates four carry signals c₀, c₁, c₂, and C₃, which it supplies to the four LUT's0-3 during an add operation. The first carry signal c₀equals the carry input C_IN, which thecarry logic3505 receives. In some embodiments, each other carry signal c_jproduced by thecarry logic3505 is derived from the propagate, generate, and carry signals from the previous stage LUT. For instance, in some embodiments, the carry signal c_jequals (P_i-1·C_i-1)+G_i-1, for cases where i can equal 1, 2, 3, or 4, and the last carry signal C_OUTequals the carry signal C₄. This carry signal C_OUTcan be output as a local carry output C_OUTL(e.g., a carry output to a tile in the same topological row) and a global carry output C_OUTG(e.g., a carry output to a tile in a different topological row) through associated buffer and/or routing circuitry (not shown).
• The sharing of the carry logic and the clustering of the logic circuits 0-3 allows the tiles in the alignedtile layout3500 to form a fast four-bit adder/subtractor. In addition, when ganged with other fast nibble wide adders/subtractors that are on the same topological row, the nibble wide adders/subtractors can form fast byte-wise adders/subtractors (as shown inFIG. 34) or other larger adders/subtractors (sixteen bit adders/subtractors, thirty-two bit adders/subtractors, etc.).
• To further speed the carry logic circuitry for larger adders/subtractors, bypass circuitry can be used to bypass the sharedcarry logic3505.FIG. 36 illustrates one such bypass circuitry. As shown in this figure, thebypass circuitry3600 includes the sharedcarry logic circuit3505, an ANDgate3610, and a two-to-one multiplexer3615. The sharedcarry logic3505 generates the carry signals (c₀, c₁, c₂, c₃, and C_OUT) based on the functions that were discussed above while describingFIG. 35. An example of the sharedcarry logic circuit3505 will be described below by reference toFIG. 39.
• When all the propagate signals generated by the logic circuits (0-3) are “1”, the AND gate produces a “1”, which directs themultiplexer3615 to output as C_OUTthe carry signal C_INthat thecarry logic3505 receives. On the other hand, when one of the propagate signals is not 1, the ANDgate3610 produces a “0”, which directs themultiplexer3615 to output the output carry signal C_OUTthat is produced by the sharedcarry logic circuit3505. Bypassing the computations of the sharedcarry circuit3505 speeds up the operation of the four-bit adder/subtractor formed by the logic and carry circuits in the alignedtile layout3500 ofFIG. 35.
• Some embodiments also use a portion of this bypass circuitry of the carry logic circuit to generate complex functions with the logic and carry circuits in the alignedtile layout3500, when these circuits are not used to implement an adder/subtractor. For instance, when all the LUT's are configured to add two one-bit values, the output S of the ANDgate3610 can be expressed as follows:
  S=(a₀⊕b₀)·(a_i⊕b_i)·(a₂⊕b₂)·(a₃⊕b₃).
  As expressed in this equation, the AND gate's output S equals the AND'ing of four XOR operations that can be performed by the four logic circuits 0-3 on their first two inputs “a” and “b”.
• Such a complex function can be used to implement a series of complex functions through NPN operations, where NPN stands for negate input (i.e., invert input), permute input, and negate output. For instance, such a function can be used to determine whether two four-bit signals are identical by inverting the four bits of one of the signals. This inversion will cause the XOR operation to produce a 1, whenever the two corresponding bits in the two signals are identical. Hence, the output of the ANDgate3610 provides the results of a four-bit comparison of two four-bit signals, when the four bits of one of the two signals are inverted, and the inverted signal is provided to the logic circuits of the aligned tile layout along with the other non-inverted signal. In such a situation, an output value of “1” for the AND gate specifies that the two four bit signals are identical, while an output value of “0” specifies a difference between the two signals. Larger comparators can be quickly created by AND'ing the outputs of the ANDgates3610 of several aligned tile layouts. For instance, a sixteen-bit comparator that can compare two sixteen-bit signals can be created by AND'ing the outputs of the ANDgates3610 of four aligned tile layouts.
• The output of the ANDgate3610 and themultiplexer3615 inFIG. 36 is fed to a sub-cycle configurable two-to-one multiplexer (not shown). Based on its configuration, this multiplexer then determines which of the two outputs it should direct to the routing fabric for routing to other circuits in the IC.
• VII. Configurable LUT that Serves as an Adder/Subtractor and Manchester Carry Chain
• FIG. 37 illustrates an example of a three-input LUT3700 of some embodiments of the invention. This LUT can be used as theLUT3205 ofFIG. 32, or the LUT. During an add or subtract operation, theLUT3700, like theLUT3205, (1) performs the actual add or subtract computation, and (2) produces the propagate and generates values that are to be used by the carry logic that will generate the next carry bit and summation.
• TheLUT3700 is implemented in complementary pass logic (CPL). In this implementation, a complementary pair of signals represents each logic signal, where an empty circle at the input or output of a circuit denotes the complementary input or output of the circuit in the figures. The LUT has three sections, acore logic section3705, a propagatesection3710, and a generate section. Thecore logic section3705 is formed by threestages3730,3735, and3740 of multiplexers that are controlled by the three input signals a, b, and c. Thecore logic section3705 generates the function f(a,b,c) computed by thelogic circuit3700.
• Given that theLUT3700 is a configurable logic circuit, the function that it computes depends on the values of configuration bits supplied to the first stage ofmultiplexers3730 of the LUT. For instance, when adding two one-bit values (i.e., computing a+b), the values of the true configuration bits are 10010110, with the most significant bit being supplied tomultiplexer input3720 and the least significant bit being supplied to themultiplexer input3725. Alternatively, the configuration bits are 01101001, when the LUT subtracts two one-bit values (i.e., computes a−b). The values of the complement configuration bits are the inverted version of their corresponding true configuration bits.
• As shown inFIG. 37, half of thefirst stage multiplexers3730 are driven by the input “a” and its complement, while the other half of thefirst stage multiplexers3730 are driven by the input “b” and its complement. The above-mentioned U.S. Patent Application entitled “Configurable IC with Interconnect Circuits that also Perform Storage Operations” (which is filed concurrently with the present application, with attorney docket number TBUL.P0022) discloses an example of a CPL-implementation of a multiplexer.
• The output of thefirst stage multiplexers3730 are supplied to thesecond stage multiplexers3735, in the manner illustrated inFIG. 37. One of the second-stage multiplexers is driven by the input signal “b”, while the other second-stage multiplexer is driven by the input signal “a”. The signals for driving the multiplexers in thefirst stage3730 and thesecond stage3735 are a mixture of the two input signals “a” and “b”, in order to balance loading and therefore delay on the signals “a” and “b”. However, in other embodiments, all the first stage multiplexers are driven only by the input “a”, while all the second stage multiplexers are driven by the input “b”, or vice versa.
• The outputs of thesecond stage multiplexers3735 are supplied to thethird stage multiplexer3740, which is driven by the input signal “c”. The output of the third stage multiplexer is the function computed by theLUT3700. This output is expressed in CPL format, i.e., in terms of the function f and its complement.
• The LUT's propagatesection3710 produces the propagate signal P and its complement. This section has two stages ofmultiplexers3750 and3755. The first stage ofmultiplexers3750 receive the lowest four significant bits of the configuration data, in the manner indicated inFIG. 37. Specifically, this figure identifies the lowest four significant configuration bits by number, and then illustrates how these four bits are supplied to thefirst stage multiplexers3750 of the propagate-generation section3710.
• Thefirst stage multiplexers3750 are driven by the input signal “b”. The output of the first multiplexer stage is supplied to amultiplexer3755 that forms the second multiplexer stage of thesection3710. Thismultiplexer3755 is driven by the input signal “a”. The output of thesecond stage multiplexer3755 represents the propagate signal P. Instead of the propagatesection3710, the LUT's of some embodiments use the propagatesection3760, which is a circuit equivalent of thesection3710 for the input configuration illustrated inFIG. 37. The output of bothsections3710 and3760 is expressed in CPL format, i.e., in terms of the propagate signal P and its complement.
• The LUT's generatesection3715 produces the generate signal G and its complement. This section includes a two-to-one CPL multiplexer that receives the input “a” and “a” along its select lines. When adding two one-bit values, the multiplexer insection3715 receives the signals “0” and “1” along its firstcomplementary pair3780 of input lines and the signals “b” and “b” along its secondcomplementary pair3785 of input lines. When subtracting two one-bit values, the multiplexer insection3715 receives thesignals1 and0 along its firstcomplementary pair3780 of input lines and the signals “b” and “b” along its secondcomplementary pair3785 of input lines. Hence, the output of this multiplexer provides the function G (which equals (a·b) when “a” and “b” are added and (a·b) when b is subtracted from a), and the complement of this function.
• FIG. 38 illustrates a three-input LUT3800 that is an optimized version of theLUT3700 ofFIG. 37. InLUT3800, the propagatesection3710 is replaced with the propagatesection3760, which was described above by reference toFIG. 37. Also, inLUT3800, the generatesection3715 has been eliminated. Instead of producing the generate signal G and its complement, theLUT3800 produces the generate signal G′ and its complement. Unlike the signal G, which equals (a·b) or (a·b), the signal G′ equals “a” while its complement equalsa.
• TheLUT3800 produces the signal G′ and its complement in such a fashion based on the following observation. As mentioned above, the carry out signal C_OUTproduced by an LCB (e.g., LCB3200) equals (P·C_in)+G, where P and G are the propagate and generate signals produced by the LCB and C_INis the carry in signal received by the LCB. The C_OUTequation can be expressed as the C_OUTequals the propagate signal when the carry in signal C_INis “1”, and equals the generate signal when the carry in signal C_INis “0”. In other words, the generate signal can be ignored unless the propagate signal is “0”.
• However, when the propagate signal is “0”, then either both “a” and “b” equal “1”, or both “a” and “b” equal “0”. When the propagate signal is “O” and the generate signal needs to be examined, the generate signal equals either “a” or “b”, both of which are equal. Accordingly, instead of computing (a·b) or (a·b) to produce a generate value G, theLUT3800 outputs a generate value G′ that equals “a” and a generate complement valueG that equals “a.”
• FIG. 39 illustrates a CPL-implementation of a four-stage Manchester carrychain3900 that can serve as the shared carry logic3605 ofFIG. 36. As shown inFIG. 39, each stage of thechain3900 includes a two-to-one CPL multiplexer (3905,3910,3915, or3920) that connects two of its four input lines to two output lines based on the two signals that it receives on its select lines.
• The multiplexer of each stage produces the carry signal of the next stage, or the output carry signal of a nibble-wide adder/subtractor, based on the propagate and generate signals generated by the LUT of the current stage and the carry out of the previous stage. For instance, thesecond multiplexer3910 in this chain produces the carry signal c₂(1) forLUT2 in a four LUT tile group (like group3500), and (2) for thethird multiplexer3910 in the carry chain. Thesecond multiplexer3910 computes the carry signal c₂as (P₁·c₁)+G₁. More specifically, thesecond multiplexer3910 sets c₂andc₂equal to c₁andc₁when the P₁equals “1”, and sets c₂andc₂equal to G′₁andG′, when the P₁equals “0”.
• Thiscarry chain3900 is referred to as a Manchester carry chain since each CPL multiplexer is formed by pass transistor logic. As mentioned above, examples of such multiplexers are described in the above-incorporated U.S. Patent Application entitled “Configurable IC with Interconnect Circuits that also Perform Storage Operations” (which is filed concurrently with the present application, with attorney docket number TBUL.P0022). One of ordinary skill will realize that other embodiments might use other types of logic to form the carry chain, such as full complex CMOS, dynamic CMOS, etc. Also, other embodiments might structure the carry chain differently. In addition,FIG. 39 illustrates thecarry chain3900 as receiving the generate signals G′ andG′, which can be produce by LUT's likeLUT3800. This carry chain, however, can also be used with LUT's likeLUT3700 that produce generate signals G andG.
• VIII. Dual Carry Chains
• Some embodiments of the invention have two carry chains in each aligned tile group to provide the IC designer maximum flexibility in arranging the data paths in the design.FIG. 40 presents a topological illustration of onesuch tile group4000. Thistile group4000 includes fourtiles4005, and fourrouting multiplexers4035,4040,4045, and4050. Eachtile4005 includes sixrouting multiplexers4010, three inputselect multiplexers4015, one three-input LUT4020. In each tile, two input select multiplexers4015 (labeled asmultiplexers 1 and 2) are HUMUX's, which receive user signals throughrouting multiplexers4035 and4040 of the tile group.
• As shown inFIG. 40, thetile group4000 also includes two carry chains, a left-to-right carry chain4025 and a right-to-leftcarry chain4030. These carry chains illustrate the direction of carry signal flow through an adder/subtractor formed by the LUT's and carry logic circuits of thetile group4000. Each carry chain receives the output of arouting multiplexer4045 or4050, which provides a local or global carry in signal. As further described below, therouting multiplexers4045 and4050 are interconnect/storage elements in some embodiments.
• As mentioned above, each LUT in some embodiments has a separate carry logic circuit, while the LUT's in other embodiments share carry logic circuits. Two carry chains can be defined in each tile group by defining a redundant set of carry logic data paths in the tile group. For instance, some embodiments establish a tile group with two carry logic chains by taking the arrangedtile layout3500 ofFIG. 35 and adding a second Manchester carrylogic3505.
• FIG. 41 illustrates one such modifiedtile layout4100. The tile layout in this figure is similar to the tile layout inFIG. 35, except that thetile layout4100 also includes (1) two Manchester carrylogic chains4105F and4105R (instead of one Manchester carry logic chain3505), (2) tworouting multiplexers4045 and4050 (instead of one routing multiplexer3510), and (3) two sets of carry in and out signals (instead of one). Thecarry logic4105F is used by the left-to-right carry chain4025, while thecarry logic4105R is used by the right-to-leftcarry chain4030. InFIG. 41, the notation F and R are used to specify the signals in the forward and reverse carrypaths4025 and4030.
• When thetile layout4100 is used to perform an adder/subtractor operation, its LUT's4120-4135 receive data and carry inputs for performing such an operation. When theforward carry chain4025 is used, the data and carry signals flow through the LUT's4120,4125,4130, and then4135. On the other hand, when thereverse carry chain4030 is used, the data and carry signals flow through the LUT's4135,4130,4125, and then4120. Accordingly, the LUT's and the inputs and outputs of the circuits inFIG. 41 are labeled to show the identity of these signals during the forward and reverse carry flows.
• The notations inFIG. 41 can be interpreted as follows.Tile layout4100 can be used to add two four-bit signals “a” and “b”, where this addition factors in a four-bit carry signal “c”. Each of the signals “a”, “b”, and “c” has abit0,bit1,bit2, andbit3. Each of the four LUT's4130-4130 always receives the same signal value in the forward and reverse flows through the LUT's. However, in the forward and reverse flows, the signal value received by each LUT is a different bit in the addition operation.
• For instance,LUT4125 is labeled as IF and2R to indicate that this circuit isLUT1 in the left-to-right adder/subtractor implementation, while it isLUT2 in the right-to-left adder/subtractor implementation. When thetile layout4100 performs an addition operation in the forward flow, the “a”, “b”, and “c” signals received by theLUT4125 are designated as a_1F, b_1F, and C_1F, to specify that these signals are the second bits in the nibble-wide add operation performed by the LUT's of thetile layout4100. Alternatively, when thetile layout4100 performs an addition operation in the reverse flow, the “a”, “b”, and “c” signals received by theLUT4125 are designated as a_2R, b_2Rand c_2R, to specify that these signals are the third bits in the nibble-wide add operation performed by the LUT's of thetile layout4100. Similarly, the propagate signal ofLUT4135 is labeled as P_3Fand P_0Rto indicate that (1) when acting as a left-to-right adder/subtractor, the propagate signal ofLUT4135 is the third propagate signal, while (2) when acting as a right-to-left adder/subtractor, the propagate signal of LUT4135 (which now is acting as LUT0) is the first propagate signal.
• As mentioned above, therouting multiplexers4045 and4050 are interconnect/storage elements, like the interconnect/storage element2700 ofFIG. 27. Similarly, in some embodiments, therouting multiplexer3510 ofFIG. 35 is also an interconnect/storage element. Using interconnect/storage elements forrouting multiplexers3510,4045, and4050 is beneficial in that it allows some embodiments to perform different portions of an adder/subtractor operation in different sub-cycles.
• For instance, to perform a thirty-two bit add operation, some embodiments can perform two sixteen bit add operations in two different sub-cycles. To do this, these embodiments can latch the carry out signal or signals associated with the addition operations for the first set of sixteen bits, in the interconnect/storage RMUX's (3510,4045, or4050) of the LUT's that perform the addition for the second set of sixteen bits, or some interconnect/storage RMUX's that are used to route the signals. While performing the addition on the second set of sixteen bits, the IC of some embodiments can simply latch the result of the addition operation on the first set of sixteen bits, or it can perform additional operations on this result in order to increase its throughput through pipelining.
• IX. Memories Embedded in and Between the Tile Layouts
• Configurable IC's typically include memory arrays for storing data used by the configurable IC. Some embodiments embed memories in the tiles of a configurable IC's tile arrangement. For example,FIG. 42 illustrates one manner of embedding amemory4205 in the layout of thetile group4000 ofFIG. 40. Thememory4200 is a 128-bit memory that can be addressed by five address bits to read or write four bits of data at a time.
• Thetile layout4200 ofFIG. 42 is similar to thetile layout4000 ofFIG. 40, except that the LUT's4020 and carrychains4025 and4030 in thelayout4000 are replaced with amemory4200 in thelayout4200. Bothlayouts4000 and4200 have four sets ofrouting multiplexers4010, four sets of inputselect multiplexers4015, and four other routing multiplexers4035-4050.
• Like the four three-input LUT's4020 inFIG. 40, thememory4205 receives the twelve bits output from the twelve inputselect multiplexers4015. However, in thelayout4200, (1) the output of the IMUX “2” in each tile and the output of therouting multiplexer4045 form a five-bit write-address bus of thememory4205, (2) the output of the IMUX “1” in each tile and the output of therouting multiplexers4050 form a five-bit read-address bus of thememory4205, and (3) the output of the IMUX “0” in each tile forms a four-bit input data bus.
• Thetile layout4200 also has a four-bit output data bus that utilizes the same four bit output data path that is used in thetile group4000 to output the four output bits of the four LUT's4020. Thetile layout4200 utilizes the output of themultiplexer4040 as the write-enable signal WE. This signal directs thememory4205 to utilize the address from the write-address bus to identify a location in the memory to write the data on the data input bus. Thetile layout4200 utilizes the output of themultiplexer4035 as a chip select signal SEL. This signal either indicates that the memory is selected for operation, or is not selected, in which case the memory can operate in a reduced power mode.
• FIG. 43 illustrates a physical layout for embedding thememory4205 in an aligned tile group, which is formed by four tiles that are aligned with each other in a manner similar to the alignedtile groups3100 and4100 ofFIGS. 31 and 41. The alignment illustrated inFIG. 43 has thememory4205 placed in the middle of the four alignedtiles4210,4215,4220, and4225, which were topologically illustrated inFIG. 42. In this embedding, thememory array4205 in the arrangement illustrated inFIG. 43 takes the place of the LUT's0-3 and shared carry logic circuits4105 inFIG. 41.
• In some embodiments, the embedding illustrated inFIG. 43 does not disrupt the routing fabric within the tiles that contain thememory4205. In these, the embedding illustrated inFIG. 43 does not utilize many or any of the configurable routing multiplexers (that are part of the configurable routing fabric of the configurable IC) in the four tiles illustrated in this figure. These unused routing multiplexers can then be used as part of the configurable routing fabric that routes signals between the configurable logic circuits of the configurable IC.
• In some architectures, the address and data signals for a memory can come from several groups of tiles.FIG. 44 illustrates onesuch architecture4400. This is a dual-ported architecture that includes two memory address/data ports4410. Eachmemory port4410 spans across two groups of eight tiles. Each port has (1) a nine-bit read address bus, (2) a nine-bit write address bus, (3) a ten-bit input data bus, and (4) a ten-bit output data bus.
• The nine-bit write address bus is formed by (1) the output of the IMUX “2” in each of the eight tiles spanned by the port, and (2) the output of one of therouting multiplexers4045 in the two groups. The nine-bit read address bus is formed by (1) the output of the IMUX “1” in each of the eight tile spanned by the port, and (2) the output of one of therouting multiplexers4050 in the two groups.
• The ten-bit data input bus is formed by (1) the output of the IMUX “0” in each of the eight tiles spanned by the port, and (2) the output of a routing multiplexer that correspond to the routing multiplexer in a computational tile (i.e., a tile with a logic circuit) that provides the carry in to the aligned tile layout. The ten-bit data output bus includes two sets of four bit lines that are each aligned with the four bit output data path used in thetile group4000 to output the four output bits of the four LUT's4020. The ten-bit data output bus also includes two bit lines that are aligned with the carry-out signal line of atile group4000 with four LUT's4020 and associated carry logic.
• These address and data lines of the dual portedarchitecture4400 allow simultaneous read and/or write operations to and/or from two different locations in a memory array, which stores 5120 bits in some embodiments. Also, in some embodiments, the two ports A and B ofFIG. 44 can operate on two different clock domains. Specifically, some embodiments can drive the circuits (e.g., the configurable routing and input-select interconnect circuits) of the two sets of tiles spanned by the two ports by two different clock signals, which potentially have different phases and/or operate at different frequencies.
• FIG. 45 illustrates one manner for establishing the dual-portedarchitecture4400 ofFIG. 44 in the tile architecture of some embodiments. Specifically,FIG. 45 illustrates a physical layout for embedding amemory4500 between four aligned tile groups in the tile architecture of some embodiments. Each aligned tile group is formed by four tiles that are aligned with each other in a manner similar to the alignedtile groups3100 and4100 ofFIGS. 31 and 41.
• The alignment illustrated inFIG. 45 has amemory4500 placed between two pairs of aligned tiles, with the top pair includingtile groups4505 and4510 and the bottom pair includingtile groups4515 and4520. The top pair oftile groups4505 and4510 provide the address and data signals for one port (e.g., port A) of thememory4500, while the bottom pair oftile groups4515 and4520 provide the address and data signals for another port (e.g., port B) of thememory4500.
• Unlike the embedding illustrated inFIG. 43, which simply takes the place of the LUT's0-3 and the shared carry logic circuits, the embedding inFIG. 45 is not within a tile layout. The embedding inFIG. 45 also requires additional wiring to route the signals from the multiplexers of the top and bottom aligned tile groups to thememory4500. However, in some embodiments, the embedding illustrated inFIG. 45 does not disrupt the routing fabric of the tiles that are on either side of thememory4500. In these embodiments, the embedding illustrated inFIG. 45 does not utilize many or any of the configurable routing multiplexers (that are part of the configurable routing fabric of the configurable IC) in the sixteen tiles illustrated in this figure. These unused routing multiplexers can then be used as part of the configurable routing fabric that routes signals between the configurable logic circuits of the configurable IC.
• X. Configurable IC and System
• Some embodiments described above are implemented in configurable IC's that can compute configurable combinational digital logic functions on signals that are presented on the inputs of the configurable IC's. In some embodiments, such computations are state-less computations (i.e., do not depend on a previous state of a value). Some embodiments described above are implemented in configurable IC's that can perform a continuous function. In these embodiments, the configurable IC can receive a continuous function at its input, and in response, provide a continuous output at one of its outputs.
• FIG. 46 illustrates a portion of aconfigurable IC4600 of some embodiments of the invention. As shown in this figure, thisIC4600 has aconfigurable circuit arrangement4605 and I/O circuitry4610. Theconfigurable circuit arrangement4605 can be any of the invention's configurable circuit arrangements that were described above. The I/O circuitry4610 is responsible for routing data between theconfigurable nodes4615 of thearrangement4605 and circuits outside of the arrangement (i.e., circuits outside of the IC, or within the IC but outside of the arrangement4605). As further described below, such data includes data that needs to be processed or passed along by the configurable nodes.
• The data also includes in some embodiments configuration data that configure the nodes to perform particular operations.FIG. 47 illustrates a more detailed example of this. Specifically, this figure illustrates aconfiguration data pool4705 for theconfigurable IC4600. This pool includes N configuration data sets (CDS). As shown inFIG. 47, the input/output circuitry4610 of theconfigurable IC4600 routes different configuration data sets to different configurable nodes of theIC4600. For instance,FIG. 47 illustratesconfigurable node4745 receivingconfiguration data sets 1, 3, and J through the I/O circuitry, whileconfigurable node4750 receivesconfiguration data sets 3, K, and N-1 through the I/O circuitry. In some embodiments, the configuration data sets are stored within each configurable node. Also, in some embodiments, a configurable node can store multiple configuration data sets so that it can reconfigure quickly by changing to another configuration data set. In some embodiments, some configurable nodes store only one configuration data set, while other configurable nodes store multiple such data sets.
• A configurable IC of the invention can also include circuits other than a configurable circuit arrangement and I/O circuitry. For instance,FIG. 48 illustrates a system on chip (“SoC”) implementation of aconfigurable IC4800. This IC has aconfigurable block4850, which includes aconfigurable circuit arrangement4605 and I/O circuitry4610 for this arrangement. It also includes aprocessor4815 outside of the configurable circuit arrangement, amemory4820, and abus4810, which conceptually represents all conductive paths between theprocessor4815,memory4820, and theconfigurable block4850. As shown inFIG. 48, theIC4800 couples to abus4830, which communicatively couples the IC to other circuits, such as an off-chip memory4825.Bus4830 conceptually represents all conductive paths between the components of theIC4800.
• Thisprocessor4815 can read and write instructions and/or data from an on-chip memory4820 or anoffchip memory4825. Theprocessor4815 can also communicate with theconfigurable block4850 throughmemory4820 and/or4825 throughbuses4810 and/or4830. Similarly, the configurable block can retrieve data from and supply data tomemories4820 and4825 throughbuses4810 and4830.
• Instead of, or in conjunction with, the system on chip (“SoC”) implementation for a configurable IC, some embodiments might employ a programmable system in package (“PSiP”) implementation for a configurable IC.FIG. 49 illustrates onesuch SiP4900. As shown in this figure,SiP4900 includes four IC's4920,4925,4930, and4935 that are stacked on top of each other on asubstrate4905. At least one of these IC's is a configurable IC that includes a configurable block, such as theconfigurable block4850 ofFIG. 48. Other IC's might be other circuits, such as processors, memory, etc.
• As shown inFIG. 49, the IC communicatively connects to the substrate4905 (e.g., through wire bondings4960). These wire bondings allow the IC's4920-4935 to communicate with each other without having to go outside of thePSiP4900. In some embodiments, the IC's4920-4935 might be directly wire-bonded to each other in order to facilitate communication between these IC's. Instead of, or in conjunction with the wire bondings, some embodiments might use other mechanisms to communicatively couple the IC's4920-4935 to each other.
• As further shown inFIG. 49, the PSiP includes a ball grid array (“BGA”)4910 and a set ofvias4915. TheBGA4910 is a set of solder balls that allows thePSiP4900 to be attached to a printed circuit board (“PCB”). Each via connects a solder ball in theBGA4910 on the bottom of thesubstrate4905, to a conductor on the top of thesubstrate4905.
• The conductors on the top of thesubstrate4905 are electrically coupled to the IC's4920-4935 through thewire bondings4960. Accordingly, the IC's4920-4935 can send and receive signals to and from circuits outside of thePSiP4900 through the wire bondings, the conductors on the top of thesubstrate4905, the set ofvias4915, and theBGA4910. Instead of a BGA, other embodiments might employ other structures (e.g., a pin grid array) to connect a PSiP to circuits outside of the PSiP. As shown inFIG. 49, ahousing4980 encapsulates thesubstrate4905, theBGA4910, the set ofvias4915, the IC's4920-4935, and the wire bondings, to form thePSiP4900. This and other PSiP structures are further described in United States Patent Application entitled “Programmable System in Package”, filed concurrently herewith with attorney docket number TBUL.P0030.
• FIG. 50 conceptually illustrates a more detailed example of acomputing system5000 that has anIC5005, which includes one of the invention's configurable circuit arrangements that were described above. Thesystem5000 can be a stand-alone computing or communication device, or it can be part of another electronic device. As shown inFIG. 50, thesystem5000 not only includes theIC5005, but also includes abus5010, asystem memory5015, a read-only memory5020, astorage device5025,input devices5030,output devices5035, andcommunication interface5040.
• Thebus5010 collectively represents all system, peripheral, and chipset interconnects (including bus and non-bus interconnect structures) that communicatively connect the numerous internal devices of thesystem5000. For instance, thebus5010 communicatively connects theIC5015 with the read-only memory5020, thesystem memory5015, and thepermanent storage device5025.
• From these various memory units, theIC5005 receives data for processing and configuration data for configuring the IC's configurable logic and/or interconnect circuits. When theIC5005 has a processor, the IC also retrieves from the various memory units instructions to execute. The read-only-memory (ROM)5020 stores static data and instructions that are needed by theIC5010 and other modules of thesystem5000. Thestorage device5025, on the other hand, is a read-and-write memory device. This device is a non-volatile memory unit that stores instruction and/or data even when thesystem5000 is off. Like thestorage device5025, thesystem memory5015 is a read-and-write memory device. However, unlike thestorage device5025, the system memory is a volatile read-and-write memory, such as a random access memory. The system memory stores some of the instructions and/or data that the IC needs at runtime.
• Thebus5010 also connects to the input andoutput devices5030 and5035. Theinput devices5030 enable the user to enter information into thesystem5000. Theinput devices5030 can include touch-sensitive screens, keys, buttons, keyboards, cursor-controllers, microphone, etc. Theoutput devices5035 display the output of thesystem5000.
• Finally, as shown inFIG. 50, thebus5010 also couples thesystem5000 to other devices through thecommunication interface5040. Examples of thecommunication interface5040 include network adapters that connect to a network of computers, or wired or wireless transceivers for communicating with other devices. One of ordinary skill in the art would appreciate that any other system configuration may also be used in conjunction with the invention, and these system configurations might have fewer or additional components.
• While the invention has been described with reference to numerous specific details, one of ordinary skill in the art will recognize that the invention can be embodied in other specific forms without departing from the spirit of the invention. For example, although numerous embodiments were described by reference to flat tile architectures, one of ordinary skill will realize that these embodiments could be implemented in other configurable IC architectures.
• Also, in some embodiments, the position of many circuits (e.g., of routing and input-select interconnects in aligned tile layouts) are topologically illustrated in the figures. The actual physical location of these circuits may be different in different embodiments. For instance, in a computation aligned tile layout that has logic circuits and routing and input-select interconnects, some embodiments position (1) the logic circuits in the center of the aligned tile layout, (2) the input-select interconnects above, below, to the right, and to the left of the centrally located logic circuits, and (3) the routing interconnects in the remaining corner areas of the aligned tile layout with other circuits.
• Many embodiments described above include input select interconnects for the logic circuits. Other embodiments, however, might not use such interconnects. Thus, one of ordinary skill in the art would understand that the invention is not to be limited by the foregoing illustrative details, but rather is to be defined by the appended claims.

Claims (33)

1. A configurable integrated circuit (“IC”) comprising:

a plurality of configurable tiles arranged in a tile arrangement, each configurable tile having a set of configurable logic circuits and a set of configurable routing circuits for routing signals between configurable logic circuits;

wherein at least a first logic circuit of a first tile has at least one direct connection with a second circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement; and

wherein said direct connection does not include any intervening interconnect circuits.

2. The configurable IC ofclaim 1, wherein a plurality of logic circuits of the first tile have at least one direct connection with at least one circuit of another tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.

3. The configurable IC ofclaim 1, wherein the second circuit is a routing circuit.

4. The configurable IC ofclaim 3, wherein the second circuit receives output of the first circuit through the direct connection.

5. The configurable IC ofclaim 1, wherein the second circuit is a logic circuit.

6. The configurable IC ofclaim 1, wherein the second circuit receives output of the first circuit through the direct connection.

7. The configurable IC ofclaim 1 further comprising configurable input-select circuits for receiving and selecting inputs for configurable logic circuits, wherein the second circuit is a configurable input-select circuit that through the direct connection receives output of the first logic circuit.

8. The configurable IC ofclaim 1, wherein each of a plurality of logic circuits of each particular tile in a set of tiles has at least one direct connection with at least one other circuit of another tile that does not neighbor the particular tile and that is not aligned horizontally or vertically with the particular tile in the tile arrangement.

9. The configurable IC ofclaim 1, wherein the direct connection does not include any intervening circuits.

10. The configurable IC ofclaim 9, wherein the direct connection includes a set of wire segments and a set of vias.

11. The configurable IC ofclaim 1, wherein the direct connection includes a set of wire segments and at least one intervening buffer circuit.

12. The configurable IC ofclaim 1, wherein the plurality of tiles includes at least300 tiles.

13. The configurable IC ofclaim 1, wherein the plurality of tiles includes at least 1000 tiles.

14. An electronic device comprising:

a configurable integrated circuit (“IC”) comprising:

a) a plurality of configurable tiles arranged in a tile arrangement,

b) each configurable tile having a set of configurable logic circuits and a set of configurable routing circuits for routing signals between configurable logic circuits;

c) wherein at least a first logic circuit of a first tile has at least one direct connection with a second circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement, and wherein said direct connection does not include any intervening interconnect circuits.

15. The electronic device ofclaim 14, wherein a plurality of logic circuits of the first tile have at least one direct connection with at least one circuit of another tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.

16. The electronic device ofclaim 14, wherein the second circuit is a routing circuit.

17. The electronic device ofclaim 16, wherein the second circuit receives output of the first circuit through the direct connection.

18. The electronic device ofclaim 14, wherein the second circuit is a logic circuit.

19. The electronic device ofclaim 14, wherein the second circuit receives output of the first circuit through the direct connection.

20. The electronic device ofclaim 14 further comprising configurable input-select circuits for receiving and selecting inputs for configurable logic circuits, wherein the second circuit is a configurable input-select circuit that through the direct connection receives output of the first logic circuit.

21. The electronic device ofclaim 14, wherein each of a plurality of logic circuits of each particular tile in a set of tiles has at least one direct connection with at least one other circuit of another tile that does not neighbor the particular tile and that is not aligned horizontally or vertically with the particular tile in the tile arrangement.

22. A configurable integrated circuit (“IC”) comprising:

a plurality of configurable tiles arranged in a tile arrangement, each particular tile having a set of configurable logic circuits, a set of configurable routing circuits for routing signals between configurable logic circuits, and a set of configurable input-select circuits for receiving inputs and configurably supplying a sub-set of the received inputs to the configurable logic circuits in the particular tile;

wherein at least a first input select circuit of a first tile has at least one direct connection with a second logic circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement, wherein the direct connection is for supplying a signal to the first input select circuit, and wherein said direct connection does not include any intervening interconnect circuits.

23. The configurable IC ofclaim 22, wherein a plurality of input select circuits of the first tile have at least one direct connection with at least one circuit of another tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.

24. The configurable IC ofclaim 22, wherein the second logic circuit is a look up table.

25. The configurable IC ofclaim 22, wherein the second logic circuit is a three input logic circuit.

26. The configurable IC ofclaim 22, wherein each of a plurality of input select circuits of each particular tile in a set of tiles has at least one direct connection with at least one other circuit of another tile that does not neighbor the particular tile and that is not aligned horizontally or vertically with the particular tile in the tile arrangement.

27. The configurable IC ofclaim 22, wherein the plurality of tiles includes at least300 tiles.

28. The configurable IC ofclaim 22, wherein the plurality of tiles includes at least 1000 tiles.

29. An electronic device comprising:

a configurable integrated circuit (“IC”) comprising:

a plurality of configurable tiles arranged in a tile arrangement,

each particular tile having a set of configurable logic circuits, a set of configurable routing circuits for routing signals between configurable logic circuits, and a set of configurable input-select circuits for receiving inputs and configurably supplying a sub-set of the received inputs to the configurable logic circuits in the particular tile;

wherein at least a first input select circuit of a first tile has at least one direct connection with a second logic circuit of a second tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement, wherein the direct connection is for supplying a signal to the first input select circuits and wherein said direct connection does not include any intervening interconnect circuits.

30. The electronic device ofclaim 29, wherein a plurality of input select circuits of the first tile have at least one direct connection with at least one circuit of another tile that does not neighbor the first tile and that is not aligned horizontally or vertically with the first tile in the tile arrangement.

31. The electronic device ofclaim 29, wherein the second logic circuit is a look up table.

32. The electronic device ofclaim 29, wherein the second logic circuit is a three input logic circuit.

33. The electronic device ofclaim 29, wherein each of a plurality of input select circuits of each particular tile in a set of tiles has at least one direct connection with at least one other circuit of another tile that does not neighbor the particular tile and that is not aligned horizontally or vertically with the particular tile in the tile arrangement.

Images