GB2325072A - Programmable logic array - Google Patents

Abstract

A programmable logic array integrated circuit has a plurality of regions 20 of programmable logic. Each region 20 includes a plurality of logic modules 22, each of which is programmable to perform any of several logic functions on input signals applied to the module. Each module 22 also includes a register for selectively registering a signal produced in or applied to the logic module. Signals for controlling the register and possibly other functions of the module 22 are selected 560 on a region-wide basis to reduce the amount of architecture control memory that each region 20 must have. Each register may be operated as either a flip-flop or a flow-through latch. The modules 22 may be interconnected in a carry chain. The main output of a first module 22 in such a chain may be fed back to the carry-in input of that module to avoid having to use another module to generate a carry-in signal for the first module in the chain.

Description

1 ' BL 2325072 LOGIC CELL FOR PROGRAMMABLE LOGIC DEVICE S

Background of the Invention

This invention relates to programmable logic array integrated circuit devices, and more particularly to improved logic cells for such devices.

Programmable logic array integrated circuit devices are well known, as is shown, for example, by Cliff et al. U.S. patent 5,260,611 and Cliff et al.

U.S. patent application No. 08/442,795, filed May 17, 1995. (Both of these documents are hereby incorporated by reference herein.) A typical device of this kind has a plurality of regions of programmable logic disposed on the device in a two-dimensional array of intersecting rows and columns of such regions. Each region (sometimes referred to as a logic array block or "LAB") includes a plurality of modules of programmable logic. Each logic module is programmable to perform any of several elementary logic functions on a plurality of inputs applied to the logic module. For example, these logic functions may include forming any logical combination of the inputs, and either registering or not registering the result to produce an output signal of the logic module.

In order to perform more complex logic functions than the logic modules can perform individually, a network of interconnection conductors is provided on the device. For example, a plurality of "horizontal" conductors may extend adjacent to each "row" of LABs, and a plurality of "vertical" conductors may extend adjacent to each "column" of LABs. A plurality of local conductors may be associated with each LAB for (1) bringing signals from the adjacent horizontal and vertical conductors into the LAB, and (2) connecting the outputs of the logic modules in the LAB to logic module inputs of the LAB. The logic module outputs of the LAB may also be connected to the horizontal'and/or vertical conductors that are adjacent to the LAB. Interconnections may also be made between horizontal and vertical conductors. Most or all of the conductor connections mentioned above may be programmable. In addition to allowing virtually any desired interconnections to be made between logic modules, the various conductors described above may also be used to bring external signals into the device and to convey output signals of the device to external circuitry.

The generalpurpose interconnection network may be supplemented by specialpurpose interconnections between adjacent or nearby logic modules. For example, to facilitate providing wide fan-in logic functions, eath logic module may have a "cascade in" input which it receives directly from a "cascade out" output of an adjacent or nearby logic module. Each logic module can logically combine the signal that results from its normal combinatorial operation with its cascade in h. p - 3 input to provide a cascade out signal. The direct cascade connections between logic modules speed the performance of wide fan-in functions.

Each logic module may be adapted to perform one place of binary addition, and to therefore produce a "carry out" signal as a secondary output signal in addition to a "sum out" signal which is the primary output of the logic module. The carry out signal of each logic module may be applied directly to a "carry in" input of an adjacent or nearby logic module. Again, these direct connections between logic modules speed the operation of a carry chain.

Logic modules may be adapted to act as various types of registers and counters (e.g., synchronous or asynchronous registers, or up or down counters, clearable counters, loadable counters, etc.).

Some prior programmable logic array devices have had extremely flexible logic modules with all of the above features programmably selectable on a logicmodule-by-logic-module basis. This requires a substantial number of programmable "architecture" memory bits for each logic module. This in turn increases the overall memory requirements of the device.

In view of the foregoing, it is an object of the present invention to reduce the architecture memory requirements of programmable logic array integrated circuit devices.

It is a more particular object of this inVention to provide logic modules and groups of logic modules (so-called regions or LABS) for programmable logic array integrated circuit devices which have considerable flexibility of use but which have reduced architecture memory requirements.

% ' p It is another more general object of this invention to provide improved logic modules and regions or LABs for programmable logic array integrated circuit devices.

Summary of the Invention

These and other objects of the invention are accomplished in accordance with the principles of the invention by providing programmable logic array integrated circuits having a plurality of regions (sometimes referred to as LABs) of programmable logic. Each region includes a plurality of logic modules, each of which is programmable to perform any of a plurality logic functions on input signals applied to the module. Each module also includes a register for selectively registering a signal produced by or applied to the module. Each module can preferably produce both a main output signal and a carry out signal. The inputs of each module preferably include a carry in signal, which is the carry out signal of another module. 20 1 Instead of providing each module with memory elements for controlling the selection of the registercontrol signals that control the register in that module, these register-control signal selections are now made on a region-wide (LAB-wide) basis for all the modules in each region (LAB) For example, the region wide register-control signal selections may include selecting signals for clocking, clearing, loading, and disabling the registers, as well as routing various moaule signals to the registers. other functions of the modules may be controlled on a similar region-wide (LAB-wide) basis. For example, a single memory element may be used to place all of the modules in a region in counter mode, and another single memory element may be b ' 0.

1 used to enable the carry in inputs of all of the modules in a region.

In accordance with another aspect of the invention, each register is made up of two, separately 5 clockable, flow-through latches connected in series. If the two latches are clocked by clock signals with an appropriate phase shift, the register functions as a flip-flop. If the clock signal of the first latch is held constant at an appropriate level, while the second latch continues to be clocked by a cyclic clock signal, the register functions as a flow-through latch. In accordance with still another aspect of the invention, the main output of selected modules can be used as the carry in inputs to those modules. A module for which this is done can better function as the first module in a carry chain because another module does not have to be used to provide the first module's carry in input and because that input is not delayed by having to come from the logic of the other module. Further features of the invention, its nature and various advantages will be more apparent from the accompanying drawings and the following detailed description of the preferred embodiments.

BZief Description of the-Drawinas

FIG. 1 is a simplified schematic block diagram of a representative portion of an illustrative embodiment of a programmable logic array integrated cil-cuit constructed in accordance with this invention.

FIG. 2 is a more specific depiction of a portion of FIG. 1.

FIG. 3 is a table showing illustrative logic performed by a portion of the FIG. 2 apparatus.

6 - FIG. 4 is a simplified schematic block diagram showing another portion of FIG. 1 in more detail.

FIG. 5 is a simplified schematic block diagram showing another feature which can be provided in circuits of the type shown in FIG. 1 in accordance with this invention.

Detailed Descri2tion of the Brefegred Embodiments The logic modules and groups of logic modules (regions or LABs) of this invention may be used in programmable logic array integrated circuit devices of the type shown in Cliff et al. U.S. patent application No. 08/672,676, filed June 28, 1996, which is hereby incorporated by reference herein. Such use of the present invention is, however, only illustrative, and the invention may be alternatively used in other types of devices.

An illustrative region or LAB 20 constructed in accordance with this invention is shown in FIG. 1 in a representative portion of a typical programmable logic array integrated circuit device 10. Device 10 is depicted as having the general type of organization of the just-mentioned Cliff et al. apparatus, although some elements and connections have been omitted from FIG. 1 in the interest of clarity and simplicity, while other elements and connections have been added to illustrate the present invention and to generalize some aspects of the device architecture. It should esecially be noted that in many cases FIG. 1 only shows a few representative conductors of various types, and that more such conductors of these types are generally provided'. For convenience in relating this specification to the just-mentioned Cliff et al.

specification, the same reference numbers are generally used for the same or similar elements in both specifications. The depicted LAB 20 is adjacent the intersection of horizontal conductors 50 and vertical conductors 60. Horizontal conductors 50 include some so-called "fast" signal lines 50e and some clock signal lines 50f. (Unlike other conductors 50 which extend only along one adjacent row of LABs 20, conductors 50e 10 and 50f may extend adjacent to all LABs on device 10.) Local conductors 40 are associated with LAB 20 for such purposes as (1) bringing signals from conductors 50 and 60 into the LAB, (2) allowing the output signal(s) of each logic module 22 in the LAB to be made available as 15 inputs to the logic modules in the LAB, and (3) allowing those same logic module output signals to be conveyed away from the LAB via conductors 50 and 60. Programmable logic connectors C'PLCs") 52 are provided for selectively bi-directionally interconnecting 20 conductors 40 and 50. PLCs 62 are similarly provided for selectively bi- directionally interconnecting conductors 40 and 60. PLCs 542 are provided for selectively applying signals on local conductors 40 to the primary inputs 30 of logic modules 22. The primary 25 output(s) 24 of each logic module is (or are) selectively applied to local conductors 40 by PLCs 26. Each logic module 22 can receive a cascade input signal 524 from an adjacent or nearby logic module and can produce a cascade output signal (also identified by the 30 relerence number 524) which is applied to another adjacent or nearby logic module. Similarly, each logic module 22 can receive a carry in input signal 526 from an adjacent or nearby logic module and can produce a

0 1 4 carry out output signal 526 which is applied to another adjacent or nearby logic module.

In addition to the above-described data signals, each logic module 22 requires a plurality of control signals, some of which are shown entering the logic modules via leads 580. Some of these control signals control the "architecture" of the logic module (e.g., whether the logic module will act as an adder stage or a counter stage). Others of these signals control certain operations of the logic module (e.g., clocking the register in the logic module, clearing that register, etc.).

In accordance with the present invention the logic module control signals are provided on a LAB-wide basis to the greatest extent that is consistent with a practical and still reasonably flexible design. (Terms like "LAB" and "LAB-wide" are used herein as convenient short-hand terms for a plurality of logic modules. The terms "region", "region-wide", and the like are used herein with a similar meaning. A "LAB" for present purposes, however, is not necessarily a LAB for all purposes. For example, a "LAB" for present purposes may be only a fraction of the logic modules served by group of local conductors 40. If local conductors 40 extend past eight logic modules 22, a LAB for present purposes may be all eight of those logic modules, or each of two groups of four of those modules, or each of two groups of three and five of those modules, etc.) In FIG. 1 the logic module control signals that are prbvided on a LAB-wide basis are those provided via leads 580. Some of these LAB-wide control signals may be derived substantially directly from programmable memory cells or function control elements U'FCEs") associated with the LAB. FCE 582 is an example of such a 1 4 1 9 - an FCE. Others of these LAB-wide control signals 580 are produced by signal decoder 570. Decoder 570 decodes signals applied to it from signal selector 560 and FCEs 572. Selector 560 is controlled by FCEs 562 to select its various output signals from among its input signals 50e, 50f, and 122. Signals 122 are selected inputs 30 to selected ones of logic modules 22.

It should be mentioned that although most logic module control signals are now LAB-wide signals 580, some logic module control signals are still determined module by module. This will be more apparent when FIG. 4 is discussed below.

FIG. 2 shows an illustrative embodiment of elements 560 and 570 in more detail, and FIG. 3 shows the decoding performed by decoder 570 in the FIG. 2 embodiment of that element. In FIG. 3 "X" indicates a "don't care" input condition, "c" indicates a clock signal (identified as 11CLW' in FIG. 2), and "c" with a bar over it indicates inversion of the clock signal.

Considering FIG. 2 now in more detail, under the control of FCEs 562, selector 560 selects from among its inputs 50e, 50f, and 122 the following signals: asynchronous clear V'MYNC CLEAR"), asynchronous load ("ASYNC LOAD"), clock enable C'CLK ENABLE"), synchronous clear ("SYNC CLEAR"), synchronous load ("SYNC LOAD"), count enable ("COUNT ENABLE"), and clock ("CLK"; also called "c" in FIG. 3). For example, selector 560 may select CLK from the signal on either ofthe two universal clock lines 50f. As another example, selector 560 may select ASYNC CLEAR from the signal on any one of the four universal fast lines 50e.

As still another example, if an R - COUNTER FCE 562 is programmed (logic 1), selector 560 selects leads 122 as 0.

4.

the sources of the SYNC CLEAR, SYNC LOAD, and COUNT ENABLE signals as follows:

1. Bottom ("D") input lead of top logic module: SYNC CLEAR; Bottom C'W') input lead of next logic module: SYNC LOAD; and Bottom ("D") input lead of bottom logic module: COUNT ENABLE.

In the particular embodiment shown, the SYNC CLEAR, SYNC LOAD, and COUNT ENABLE signals are only meaningful when R COUNTER is logic 1. Thus these counter-function-related signals always come from leads 122, thereby effectively "stealing" the D inputs from the three logic modules identified in the list which is immediately above. This stealing is acceptable because when counting, a logic module needs no more than two. primary inputs 30 (i.e., feedback from its register and one place of a count interval value). On the other hand, selector 560 may typically select signals like ASYNC CLEAR, ASYNC LOAD, and CLOCK ENABLE from any of leads 50e and 122. The CLK signal is typically selected from any of leads 50f and 122. Other sets of choices are of course possible (e.g., selecting any of ASYNC CLEAR, ASYNC LOAD, CLOCK ENABLE, and CLK from any of leads 50e, 50f, and 112).

As shown in FIG. 3, when the eight input signals shown in the upper part of that FIG. have the values specified in a particular column, the six output signals of decoder 570 have the values shown in the loWer part of that same column. For example, when the ASYNC CLEAR input to decoder 570 is logic 1, the decoder produces the following six output signal values: CLEAR = 1, LOAD = 0, DISABLE = 0, NORMAL = 0, MASTER = 1, and SLAVE = 1.

a 11 FIG. 4 shows in more detail an illustrative embodiment of a representative logic module 22 from FIG. 1. As shown in FIG. 4, logic module 22 includes a modified four-input look-up table 200. The modifications to a basic four-input look-up table are to enable one logic module to perform one place of binary arithmetic and to accordingly produce both a sum out digit (on lead 202) and a carry out digit (on lead 526). This is described in more detail in Cliff et al. U.S. patent 5,224,581, which is hereby incorporated by reference herein. Without going into all the detail provided in the last-mentioned patent, the four-input look-up table is divided into two threeinput look- up tables 530a and 530b. Each of tables 530 is programmable to produce an output which is any logical combination of the three inputs applied to that table. Normally (when LAB-wide R-CARRY FCE 582 is programmed to be logic 0), the three inputs to each of tables 530 are the A, B, and C inputs 30 to the logic module. The fourth (D) input 30 to the logic module controls PLC 532 to select the output signal of one of tables 530 as the final output signal 202 of the lookup table structure.

On the other hand if LAB-wide R-CARRY FCE 582 is programmed logic 1 (to make it possible for each logic module 22 in the associated LAB 20 to perform a respective place of binary arithmetic), then PLC 528 selects the carry in signal on lead 526 as the third input to table 530a (in place of normal input C). In this case table 530a is programmed to produce an output signal which is the sum of the A, B, and carry in inputs, and input D is used to control PLC 532 to select the output of table 530a as the sum out output signal 202 of the logic module. At the same time, the 1 a.

1 carry in signal is used in table 530b to select the carry out signal on lead 526 from two combinations of the A and B inputs produced by table 530b. Thus when R-CARRY is logic 1, logic module 22 performs one place of binary arithmetic on its A and B inputs, producing both a sum out signal on lead 202 and a carry out signal on lead 526. In accordance with this invention, the selection of the arithmetic carry mode of operation is preferably made on a LAB-wide basis by one R-CARRY FCE 582 for each LAB.

Downstream from look-up table section 200, logic module 22 includes a cascade logic section, a principal component of which is NAND gate 538. If it is desired for logic module 22 to respond to a cascade input signal 524 (applied directly from the cascade out lead 524 of another nearby logic module), FCE 534 is programmed logic 0. FCE 534 is an example of an architecture feature which is preferably still controlled on a logicmodule-by-logic-module basis in accordance with this invention. Therefore, as shown in FIG. 4, each logic module 22 preferably has its own separately programmable FCE 534. Programming FCE 534 logic 0 enables OR gate 536 to apply the cascade in signal it receives via lead 524 to NAND gate 538 for logical combination with look-up table output 202. (The feedback loop including transistor 539 is a level restoring circuit.) The output signal of NAND gate 538 is applied to cascade out lead 524 via inverter 540.

Downstream from the cascade logic section, thh next section of logic module 22 controls whether the output of the cascade logic section or the C input to the logic module is applied to the register 210 of the logic module. The normal condition of the logic module is for the output of the cascade logic section 1.

- 13 to be applied to the register. This occurs when the NORMAL output signal of decoder 570 is logic 1, thereby enabling transistor 204a. Alternatively, when the LOAD output signal of decoder 570 is logic 1, transistor 204b is enabled to pass the logic module's C input. The LOAD signal is asserted when doing a synchronous or asynchronous loading of the register (as in a loadable counter operation) or when using the logic module in a packed register mode of operation (e.g., when using the logic module to produce both a combinatorial output (based on any or all of inputs A, B, and D) and a registered output (based on input C)). The packed register mode of operation (sometimes called the lonely register mode of operation) allows the register of the logic module to continue to perform some function if desired (i.e., by serving as a register for the C input of the logic module), while the remainder of the logic module is performing a separate combinatorial function. The packed register mode of operation is associated with the right-hand column in FIG. 3 (R_PACKED_REG FCE 572 programmed logic 1) - Downstream from the NORMAL/LOAD section is the CLEAR/DISABLE section. The NORMAL, LOAD, CLE.A.R, and DISABLE signals are mutually exclusive of one another. That is, only one of these signals can be asserted (logic 1) at any one time. When the CLEAR output signal of decoder 570 is logic 1, transistor 504a is enabled, thereby tying the data input of the logic module's register to VSS (logic 0) This will clear the register. on the other hand, when the DISABLE output signal of decoder 570 is logic 1, the data input of the register is tied to the register's data output. This prevents the register from changing state and thereby disables it.

The next section of the circuitry shown in FIG. 4 is the actual register 210 of the logic module.

This circuitry has two modes of operation: (1) an edgetrigger mode and (2) a flow-through mode. The register circuitry basically comprises two flow-through latches connected in series. The left-hand latch (including forward inverter 510a) is sometimes referred to as the master stage, while the right-hand latch (including forward inverter 510b) is sometimes referred to as the slave stage. Transistors 511a and 512a comprise a weak reverse inverter from the output of inverter 510a back to its input, thereby creating the master latch. Transistors 511b and 512b similarly comprise a weak reverse inverter from the output of inverter 510b back to its input, thereby creating the slave latch. Transistor 513 is a level restorer for the output of inverter 510a.

In the edge-trigger mode of operation (e.g., third from right-hand column in FIG. 3), the MASTER signal is the inverse of the CLK signal applied to decoder 570, while NMASTER (the inverse of MASTER) and SLAVE are equal to the CLK signal. Accordingly, the master section of register 210 latches in the data signal on lead 505 in response to the rising edge of the CLK signal. The slave section also begins to output that same data signal at the same time, and the slave section latches in that data signal on the falling edge of the CLK signal. This provides conventional flip-flop operation of register 210.

In the flow-through mode of operation (second from right-hand column in FIG. 3; R_LATCH FCE 572 programmed logic 1), the MASTER signal is tied to WC - (by decoder 570). The NMASTER signal is therefore VSS. Decoder 570 continues to supply CLK as the SLAVE signal. With the MASTER signal tied to VCC, the master section merely passes through the data applied to it.

The slave section continues to operate as a flowthrough latch (as described above). This gives the overall register 210 the operating characteristics of a flow-through latch.

The final section of the circuitry shown in FIG. 4 is the output signal switching section. The output signal of register 210 is inverted by inverter 518 and then applied to one input of each of PLCs 206 and 208. The other input to each of PLCs 206 and 208 is the combinatorial output of NAND gate 538. Each of PLCs 206 and 208 is controlled by a respective FCE 207 or.209 to pass either of the applied signals. Thus either of the two outputs 24a and 24b of the logic module can be derived from either the registered or combinatorial signals of the logic module. (Above- mentioned application No. 08/672,676 shows how two outputs 24a and 24b from a logic module may be separately connected into the interconnection conductor network of the device.) FCEs 207 and 209 are preferably like cascade select FCE 534 in that they are separately provided for each logic module.

Returning now to further consider some of the input conditions shown in FIG. 3, it will be apparent that assertion of ASYNC CLEAR (first col= in FIG. 3) immediately clears register 210. Assertion of ASYNC LOO (second column in FIG. 3) immediately loads register 210 with the data signal on lead 505. Assertion of SYNC CLEAR (fourth column in FIG. 3) clears register 210 on the next rising edge of the CLK signal. Assertion of SYNC LOAD (fifth column in 1 - 16 FIG. 3) loads register 210 with the data signal on lead 505 on the next rising edge of the CLK signal. If the associated LAB 20 is configured as a counter (associated R - COUNTER FCE 562 is programmed logic 1, SYNC-CLEAR = 0, AND SYNC-LOAD = 0), the counter only counts up or down while COUNT ENABLE is asserted (seventh column in FIG. 3). (The direction of counting (i.e., whether up or down) is controlled by the programming of the look-up tables 200 in the LAB. The feedback path for feeding the contents of register 210 back to the inputs of the logic module for counting is through the local interconnect (24, 26, 40, 542, 30) of the LAB.) The CLOCK ENABLE signal controls the availability of the CLK signal for the synchronous clear, synchronous load, and counter operations. The last column in FIG. 3 represents the packed register or lonely register mode (R_PACKED_REG FCE 572 programmed logic 1).

FIG. 5 illustrates a feature of the invention which allows the carry in signal of the first logic module 22 in each LAB 20 to be either the carry out signal from the last logic module in an adjacent or nearby LAB or the primary output 24 from that same first logic module. This selection is performed for each LAB by an associated PLC 590, which is controlled by an associated R_Q_FEEDBACK FCE 592. This feature allows the first logic module of any LAB to be used as the first stage of a counter without having to use another logic module to supply an artificial carry in signal to that first logic module. In addition to saving a logic module, this expedient increases counter speed because the carry in signal for the first logic module in the counter does not have to be generated by the look-up table 200 (or, more precisely, 530b) of another logic module.

The PLCs mentioned throughout this specification (which includes the appended claims) can be implemented in any of a wide variety of ways. For example, each PLC can be a relatively simple programmable connector such as a switch or a plurality of switches for connecting any one of several inputs to an output. Alternatively, each PLC can be a somewhat more complex element which is capable of performing logic (e.g., by logically combining several of its inputs) as well as making a connection. In the latter case, for example, each PLC can be product term logic, implementing functions such as AND, NAND, OR, or NOR.

Examples of components suitable for implementing PLCs.. are EPROMs,. EEPROMs, pass transistors, transmission gates, antifuses, laser fuses, metal optional links, etc. As has been mentioned, the components of PLCs can be controlled by various, programmable, function control elements ("FCEs"), which are not always shown separately in the accompanying drawings. (With certain PLC implementations (e.g., fuses and metal optional links) separate FCE devices are not required, so that in those cases any depiction of FCE devices in the accompanying drawings merely indicates that the PLCs are programmable.) FCEs can also be implemented in any of several different ways. For example, FCEs can be SRAMs, DRAMs, first-in first-out ("FIFO") memories, EPROMs, EEPROMs, function control registers (e.g., as in'Wahlstrom U.S. patent 3,473,160), ferro-electric memories, fuses, antifuses, or the like. From the various examples mentioned above it will be seen that this invention is applicable both to one-time-only programmable and reprogrammable devices.

It will be understood that the foregoing is only illustrative of the principles of the invention and that various modifications can be made by those skilled in the art without departing from the scope and spirit of the invention. For example, the number of logic modules in a LAB can bevaried if desired. Similarly, the number of inputs 30 to each logic module can be varied if desired, as can the numbers of any of the various kinds of interconnection conductors 40, 50, 10 and 60. It will also be understood that terms like "row" and "column". "horizontal" and "vertical", "top" and "bottom", "left" and "righC, and other similar directional or orientational characterizations are entirely arbitrary and are employed only as relative terms for convenience herein. These terms are not intended to have any absolute or fixed meaning or to limit the scope of the claims to any particular device orientations or directions.

4 19

Claims (2)

1. A programmable logic array integrated circuit comprising: a plurality of modules of programmable logic, each of said modules being programmable to produce any of a plurality of logical functions of a plurality of input signals applied to the module to produce a main output signal and, depending on the logical function, a carry out signal, the input signals of each module including a carry in signal which is generally the carry out signal of a preceding module in a carry chain; and feedback circuitry associated with a predetermined one of said modules which can be a first module in a carry chain for applying the main output signal of said module to the carry in input of said module.

2. The apparatus defined in claim 1 wherein said feedback circuity comprises; a programmable selector for applying either the main output signal of said predetermined one of said modules or the carry out signal of another one of said modules to said carry in input of said predetermined one of said modules.