Abstract: The present disclosure includes apparatuses and methods for computing reduction and prefix sum operations in memory. A number of embodiments include processing circuitry configured to compute a reduction operation on data stored in a group of memory cells by splitting the data into a plurality of elements, copying each of the plurality of elements into elements that are wider than before being copied, and performing a logical operation associated with the reduction operation on each of the copied elements.

Abstract: The present invention provides a means for optimization and re-use of hardware in the implementation of Viterbi and Turbo Decoders using carry save arithmetic. Successful provision for each target application requires that two main issues be confronted. These are: merging the computation of summation terms (a2?b2+c2) with (x2+y2+z2); and (a3+b3?c3) with (x3+y3+z3); implementing an efficient method of computing (a4?b4?c4); and merging this computation with (x4+y4+z4). The invention solves both of these issues and successfully merges the Viterbi instructions with a complete reuse of the hardware that is required for the implementation of Turbo instructions. The hardware required by both classes of instructions is optimized by efficiently employing carry save arithmetic.

Abstract: Processing circuitry is provided for performing a shift-round-and-accumulate operation. The operation comprises shifting an input value to generate a shifted value using shifting circuitry, adding the shifted value to an accumulate value using adding circuitry, and performing rounding by adding a rounding value to the sum of the shifted value and the accumulated value using the adding circuitry. The same adding circuitry is used to perform both the addition of the shifted value and the accumulated value and the addition of the rounding value in the same processing cycle.

Abstract: A predictive adder produces the result of incrementing and/or decrementing a sum of A and B by a one-bit constant of the form of the form 2k, where k is a bit position at which the sum is to be incremented or decremented. The predictive adder predicts the ripple portion of bits in the potential sum of the first operand A and the second operand B that would be toggled by incrementing or decrementing the sum A+B by the one-bit constant to generate and indication of the ripple portion of bits in the potential sum. The predictive adder uses the indication of the ripple portion of bits in the potential sum and the carry output generated by evaluating A+B to produce the results of at least one of A+B+2k and A+B?2k.

Abstract: A shared parallel adder tree for executing multiple different population count operations on a single datum includes a number of carry-save adders (CSAs) and/or half adders (HAs), arranged in rows, where certain CSAs and HAs are dedicated to a single population count operation, while other CSAs and HAs are shared among two or more population count operations. The datum is applied to the first row in the tree. Partial sums of the number of ones at various locations within the tree are routed to certain CSAs and/or HAs “down” the tree to propagate the particular population count operations. Carry-propagate adders generate at least a portion of the final sum of the number of ones in certain population count operations. An “AND” operation on a particular number of the bits in the datum provides the high order bit of the resulting sum of the particular population count operation.

Abstract: A system for fast determination of a horizontal minimum of multiple digital values including a difference circuit and a compare circuit. The difference circuit may include first and second adders in which the first adder compares upper bits of a first digital value with upper bits of a second digital value and provides a first carry output and a propagate output. The second adder compares lower bits of the first digital value with lower bits of the second digital value and provides a second carry output. The compare circuit determines whether the first digital value is greater than the second digital value based on the carry and propagate outputs. Multiple difference circuits may be used to compare each of multiple digital values with every other digital value to provide corresponding compare bits, which are then used to determine a minimum one of the digital values and its corresponding location.

Abstract: Low density parity check (LDPC) decoding can be mapped to a class of DSP instructions called MINST. The MINST class of instructions significantly enhance the efficiency of LDPC decoding by merging several of the functions required by LDPC decoders into a single MINST instruction. This invention is an efficient implementation of the MINST class of instructions using a configurable three input arithmetic logic unit. The carry output results of the three input arithmetic logic unit enable permit boundary decisions in a range determination required by the MINST instruction. The preferred embodiment employs 2's complement arithmetic and carry-save adder logic. This invention allows reuse of hardware required to implement MAX* functions in LDPC functions.

Abstract: The present invention provides a means for optimization and re-use of hardware in the implementation of Viterbi and Turbo Decoders using carry save arithmetic. Successful provision for each target application requires that two main issues be confronted. These are: merging the computation of summation terms (a2?b2+c2) with (x2+y2+z2); and (a3+b3?c3) with (x3+y3+z3); implementing an efficient method of computing (a4?b4?c4); and merging this computation with (x4+y4+z4). The invention solves both of these issues and successfully merges the Viterbi instructions with a complete reuse of the hardware that is required for the implementation of Turbo instructions. The hardware required by both classes of instructions is optimized by efficiently employing carry save arithmetic.

Abstract: A logic circuit has a first logic element (“LE”) including a first lookup table (“LUT”), where the first LUT is operable to produce a carry from a first set of bits of at least two numbers. The logic circuit also has a second LE including a second LUT, where the second LUT is operable to produce a sum from a second set of bits of the at least two numbers. The second LE also includes an adder coupled directly to the first LUT and coupled to the second LUT, where the adder is operable to add the carry and the sum. The at least two numbers may be three numbers, but the logic circuit includes a set of connections operable to programmably interconnect selected inputs so that the logic circuit is operable to add only two numbers. The logic circuit may be incorporated in a programmable logic device.

Abstract: A variable node processing unit with N+1 inputs, having at least a first bank of two-input adders and a separate last bank of two-input adders, where the banks of adders are disposed in series.

Abstract: A method for generating partial sums from at least four multiple-digit sequences in a computing device includes partitioning the multiple-digit sequences into at least a first set of multiple-digit sequences and a second set of multiple-digit sequences. The method also includes generating at least one auxiliary set of multiple-digit sequences. The auxiliary set includes digits copied from respective digit positions of multiple-digit sequences in the first and second sets. The method further includes replacing the copied digits in the first and second sets by zeros to obtain a first altered set and a second altered set, respectively, of multiple-digit sequences each comprising multiple segments separated by the replaced zeros.

Abstract: The adder-subtractor includes a pre-processing circuit which divides three inputted terms into a mantissa having an exponent of maximum value, mantissa having an exponent of intermediate value and mantissa having an exponent of minimum magnitude and outputting a mantissa obtained by right-shifting the mantissa having the exponent of intermediate value and the mantissa having the minimum exponent of 2n+3 bits and adjusting digits and the mantissa having the maximum exponent, which reduces the mantissas from three to two terms, which carries out addition on the mantissas of the two terms, a normalization circuit which makes left shift so that the most significant bit becomes 1, a rounding circuit which uses an (n+3)th bit from the most significant bit as a new sticky bit, takes logical OR with the lower bits and performs rounding and an exponent operation unit which outputs a final exponent.

Abstract: The core of this invention is the application of a fast comparison circuit to the problem of address translation. Traditional implementations generate the virtual address and the physical address in series. This invention generates the physical address and virtual address simultaneously. A bitwise operation on the base address, the offset address and each stored virtual address determines whether the base address and offset address sum equals the virtual address without requiring a carry propagate. Circular addressing is implemented in the match determination by masking bits corresponding to the circular address limit.

Abstract: Adder/rounder circuitry for use in a programmable logic device computes a rounded sum quickly, and ideally within one clock cycle. The rounding position is selectable within a range of bit positions. In an input stage, for each bit position in that range, bits from both addends and a rounding bit are processed, while for each bit position outside that range only bits from both addends are processed. The input stage processing aligns its output in a common format for bits within and outside the range. The input processing may include 3:2 compression for bit positions within the range and 2:2 compression for bit positions outside the range, so that further processing is performed for all bit positions on a sum vector and a carry vector. Computation of the sum proceeds substantially simultaneously with and without the rounding input, and rounding logic makes a selection later in the computation.

Abstract: A digital logic circuit includes at least one stage. Each stage includes sum logic, combinatorial logic, and carry chain logic. The sum logic is configured to generate a first sum signal from a first set of three input signals. The combinatorial logic includes a carry generation portion and a sum generation portion. The carry generation portion is configured to generate a first carry signal from a second set of three input signals. The sum generation portion is configured to generate a second sum signal from the first sum signal and the first carry signal. The carry chain logic is configured to process the first sum signal, the second sum signal, and a carry-in signal to generate a carry-out signal and a third sum signal.

Abstract: Provided are a simplified 4:2 carry save adder (CSA) cell and a 4:2 carry save adding method. The 4:2 CSA cell is formed of an odd detector and first through sixth switches through logic optimization. The odd detector generates an XOR of the first through fourth input signals, outputs the XOR as an odd signal, generates an XOR of the first and second input signals, and outputs the XOR as a first XOR signal. The first switch outputs the third input signal as a carry output signal in response to the first XOR signal. The second switch outputs the first input signal as the carry output signal in response to an inverted first XOR signal. The third switch outputs the carry input signal as a carry signal in response to the odd signal. The fourth switch outputs the fourth input signal as the carry signal in response to an inverted odd signal. The fifth switch outputs an inverted carry input signal as a sum signal in response to the odd signal.

Abstract: Performing a sum of numbers operation in a variable bit-length environment of a processor in response to a summation instruction, comprising a) adding a least significant portion (LSP) of a first number to a LSP of another number from a plurality of numbers, wherein the sum is stored in a first storage location; b) incrementing an overflow counter if a carry is generated by adding the LSPs of the two numbers; c) adding a LSP of a next number from the plurality of numbers to the sum stored in the first storage location, wherein the resulting sum is stored back into the first storage location; d) incrementing the overflow counter if a carry is generated by adding the LSP of the next number to the sum in the first storage location; e) performing steps c) and d) until each of the LSPs of the plurality of numbers has been added.

Abstract: A data processing apparatus for summing data values includes: a plurality of adder logic stages arranged in parallel; a control logic, in response to a request to sum two data values, to forward portions of the two data values to respective ones of the plurality of adder logic stages, each of the plurality of adder logic stages performing a carry propagate addition of the received portions to generate an intermediate sum, a propagate value and a carry; and further logic stages for combining the intermediate sums, carries and propagate values to produce a sum of the two data values. The control logic, further in response to a request to add a third data value to the sum before the further logic has completed sum, forwards portions of the third data value to respective ones of the plurality of adder logic stages, feedbacks the intermediate sums, and selectively feedbacks a carry generated from a preceding adder logic stage.

Abstract: Logic circuits that support the addition of three binary numbers using hardwired adders are described. In one embodiment, this is accomplished by using a 3:2 compressor (i.e., a Carry Save Adder method), using hardwired adders to add the sums and carrys produced by the 3:2 compression, and sharing carrys data calculated in one logic element (“LE”) with the following LE. In such an embodiment, with the exception of the first and last LEs in a logic array block (“LAB”), each LE in effect lends one look-up table (“LUT”) to the LE below (i.e., the following LE) and borrows one LUT from the LE above (i.e., the previous LE). The LUT being lent or borrowed is one that implements the carry function in the 3:2 compressor model. In another aspect, an embodiment of the present invention provides LEs that include selectors to select signals corresponding to the addition of three binary numbers mode.

Abstract: The invention relates to an adder for adding at least four bits of the same significance w, said adder having a first number of inputs for receiving the bits of the same significance w that are to be added and a number of outputs, the bits to be added being applied to the inputs in presorted form, and the adder adding the bits while taking account of the presorting. The invention also provides an adding device for adding at least four bits of equal significance and a corresponding method.

Abstract: The present invention relates to a method and system for providing a variable width, at least six-way addition instruction in a processor. The method includes decoding an instruction as a variable width, at least six-way addition instruction, where the variable width, at least six-way addition instruction includes a plurality of operands. The method also includes adding the plurality of operands to obtain a plurality of sums. The method further includes outputting the plurality of sums and optionally storing carry results from the adding operation.

Abstract: A function block allows a multiplier and a multi-input multiplexer to be realized with a small number of blocks. A logical function generator generates a logical output signal from first to fourth logical inputs thereof according to a logical function selected from a plurality of 4-input/1-output logical functions depending on configuration data. A 4-2 carry block generates a 4-2 carry output from the second to fourth logical inputs. A first signal is generated from at least the logical output, a second signal from at least the first logical input, a third signal from at least a 4-2 carry input signal, and a fourth signal from at least the 4-2 carry input signal. A multiplexer selects one of the second and third signals depending on the first signal to produce a carry output signal. An exclusive OR circuit produce an exclusive-ORed result from the logical output and the fourth signal.

Abstract: A 3-input adder/subtractor unit, having a first input for receiving a first operand A, a second input for receiving a second operand B, and a third input for receiving a third operand C. An add/subtract unit includes a control input for receiving a user-specified opcode, a first 3-to-2 compressor for receiving a respective least significant bit of said operands or its complement, and a Half Adder coupled to the first 3-to-2 compressor and responsive to an output thereof and to said opcode for outputting a least significant bit of a sum equal to A+B+C or A+B?C or A?B+C or A?B?C. A plurality of 3-to-2 compressors each in respect of remaining bits of the operands receive a respective bit of the operands or its complement, and a 2-input adder coupled to the 3-to-2 compressors receive respective carry and save outputs thereof and computing respective bits of the sum apart from the least significant bit.

Abstract: An addition circuit for producing a sum of four redundant binary numbers includes a 4:2 compression adder for receiving each of the operand fields of the four redundant binary numbers, and producing a first sum field and a first carry field therefrom. The addition circuit further includes a 4:3 compression adder for receiving each of the sparse carry-save fields of the four redundant binary numbers, and producing a second sum field therefrom. The addition circuit also includes a 3:2 compression adder for receiving the first sum field, the first carry field and the second sum field, and producing a third sum field and a second carry field therefrom. The third sum field and the second carry field are the final results from addition of the four redundant binary numbers.

Abstract: Apparatus for determining a value, a sign and an overflow status of an addition of at least three n-bit data inputs. The apparatus comprising: a first adder, for adding the at least three n-bit data inputs, to provide a first output having at least 2n bits; a second adder for adding a portion of bits of the first output, the second adder being operable to add a plurality of m-bit addends, m being smaller than or equal to n. The apparatus further comprising at least two electronic-circuits, operatively associated with the first adder and the second adder. The first adder, the second adder and the at least two electronic-circuits are constructed and designed to obtain the value, the overflow status and a sign of the addition of the at least three data inputs, using predetermined parity rules being associated with a parity characteristic of the at least three data inputs.

Abstract: A carry ripple adder contains five first inputs for accepting five input bits having equal significance w that are to be summed and two second inputs for accepting two carry bits having the significance w. It also contains an output for a sum bit having the significance w and two outputs for two carry bits having the significances 2w and 4w.

Abstract: A pipelined processor such as an averaging filter including at least one subtractor section and at least one adder section. Both of the subtractor section and the adder section have a plurality of adder logic units. In comparison to the conventional processor, the processor of the present invention is streamlined by the application of one or more of three techniques. First, there is the interleaving approach where the subtractor section and the adder section are interleaved with one another. Second, there is the one delay feedback approach where the adder section includes a one delay feedback for each of the adder logic units. Third, there is the delay enable signal output approach where the averaging filter includes a delay enable signal output for each of the adder logic units of the adder section.

Abstract: The microarchitecture of the arithmetic unit includes two cascaded N bit adders to provide an N bits result in an accumulator. The arithmetic unit also includes a carry save adder, followed by an adder, which, along with the accumulator, are extended to N+1 bits. A circuit for determining the output carry value associated with the result is also provided.

Abstract: The relationship between a sum of applied address operands and a matching virtual page number is exploited to minimize the adder size required for fast number comparison. In one embodiment, variably-sized addresses are accommodated by augmenting a portion of the applied address operands to ensure easy access to potential carry bits. A comparator is used for each virtual page number stored in a translation look-aside buffer to quickly determine whether that virtual page number matches the applied address operand sum.

Abstract: A method of performing in-place arithmetic, particularly addition and subtraction, on numbers stored in respective consecutive rows of an array processor that has two tags registers. In a first machine cycle per bit, results of logical operations are stored in the tags registers, and the tags registers are shifted to align the intermediate results with other rows. In a second machine cycle per bit, results of further logical operations are stored in the tags registers, and the tags registers are shifted back to align the new intermediate results with the original rows.

Abstract: A computing system includes a plurality of full adders that each receives a bit-wise inversion of a bit of a first data, a bit of a second data, and a bit of a third data, respectively, and provides a sum output and a carry output. An exclusive-OR logic module receives the sum output of a first of the plurality of full adders and a carry output of a second of the plurality of full adders and provides an exclusive-OR output. An AND logic module has a plurality of inputs and an AND output, wherein the exclusive-OR output is electrically connected to one of the plurality of inputs of the AND logic module, and the AND output provides a signal that indicates whether the first data equals the sum of the second data and third data.

Abstract: Methods of adding and subtracting sets of binary numbers using an associative processor. The inner loop over corresponding bits of the operands is executed in only three machine cycles. Only the carry bit of each loop iteration is carried forward to the next loop iteration. At most five logical operations are used per loop iteration for addition, and at most seven logical operations, of which at most five are binary logical operations, are used per loop iteration for subtraction. In each loop iteration, the second input bit is a direct or indirect argument of at most three logical operations in addition, and of at most four logical operations in subtraction. Each loop iteration includes at least one OR operation and at most two XOR operations.

Abstract: An adder or an integrated circuit including an adder, includes a hierarchical carry-select split adder capable of operating in a split mode of operation when a mode select input takes on a first state. It is also capable of operating in a hierarchical carry-select mode of operation when the mode select input takes on a second state.

Abstract: There is disclosed a circuit for determining the number of Logic 1 bits in a group of N data bits.

Abstract: A pipelined four-to-two compressor includes sequential elements with embedded logic. One sequential element is a flip flop with complementary outputs that includes a master stage with embedded logic, a sensing stage, and one or more slave stages. The flip flop operates in a pre-charge state and an evaluate state. During the pre-charge state when a clock signal is low, the flip flop pre-charges internal keeper nodes to a high value. When the clock signal transitions high, the flip flop enters an evaluation state and one of the internal keeper nodes evaluates to a low value. Keeper nodes can also be dynamic flip flop outputs that pre-charge each clock cycle. Another flip flop with embedded logic receives the dynamic output, applies further logic, and provides a static output.

Abstract: An addition circuit for producing a sum of four redundant binary numbers includes a 4:2 compression adder for receiving each of the operand fields of the four redundant binary numbers, and producing a first sum field and a first carry field therefrom. The addition circuit further includes a 4:3 compression adder for receiving each of the sparse carry-save fields of the four redundant binary numbers, and producing a second sum field therefrom. The addition circuit also includes a 3:2 compression adder for receiving the first sum field, the first carry field and the second sum field, and producing a third sum field and a second carry field therefrom. The third sum field and the second carry field are the final results from addition of the four redundant binary numbers.

Abstract: A four-input to two-output adder is disclosed. The four-input/two-output adder includes a sum-lookahead full adder and a modified full adder. The sum-lookahead full adder includes an XOR3 block and an AXOR block for receiving a first input, a second input, a third input, and an input from a forward adjacent adder to generate a first sum signal and a sum-lookahead carry signal, respectively. The modified full adder includes an XOR2 block and a MUX2 block for receiving the first sum signal from the sum-lookahead-full adder, a fourth input, and a sum-lookahead carry signal from a backward adjacent adder to generate a second sum signal and a carry signal, respectively.

Abstract: A five-input/two-output binary adder is disclosed. The five-input/two-output adder includes five inputs and two outputs. Four levels of XOR logic gates are coupled between the five inputs and the two outputs for combining values received at the five inputs and generating a sum value and a carry value at the outputs.

Abstract: A Wallace tree structure such as that used in a DSP is arranged to sum vectors. The structure has a number of adder stages (365, 370, 375), each of which may have half adders (300) with two input nodes, and full adders (310) with three input nodes. The structure is designed with reference to the vectors to be summed. The number of full- and half-adders in each stage and the arrangement of vector inputs depends upon their characteristics. An algorithm calculates the possible tree structures and input arrangements, and selects an optimum design having a small final stage ripple adder (380), the design being based upon the characteristics of the vector inputs. This leads to reduced propagation delay and a reduced amount of semiconductor material for implementation of the DSP.

Abstract: A method and circuit for summing multiples vectors is disclosed. The method includes receiving a set of input vectors and generating a set of decoded summation vectors. Each of the set of decoded summation vectors indicates the value of at least a portion of the vector sum. The method further includes generating a set of decoded carry vectors. Each carry vector is used to select the summation vector for an adjacent portion of the vector sum from a set of preliminary summation vectors. In one embodiment, the method further includes counting the number of high bits in each bit position of the input vectors and generating decoded high bit count vectors based upon the counting to facilitate the generation of decoded summation vectors. In one embodiment, the set of preliminary vectors includes an initial preliminary summation vector and a set of adjacent summation vectors. In this embodiment each adjacent summation vector is achieved with a 1-bit rotation of the preceding adjacent summation vector.

Abstract: A data processing system for the compression and decompression of data using Differential Pulse Code Modulation, and optimized for fast execution using a parallel processing DSP such as the Texas Instruments TMS320C8X family. Decompression is speeded up over the methods known in the art by combining the VLC codes and the additional interval bits into one code, which is then used as an index into a Look Up Table that yields the final result in one step instead of the multiple operations required by current implementations.

Abstract: A computing system includes a plurality of full adders that each receives a bit-wise inversion of a bit of a first data, a bit of a second data, and a bit of a third data, respectively, and provides a sum output and a carry output. An exclusive-OR logic module receives the sum output of a first of the plurality of full adders and a carry output of a second of the plurality of full adders and provides an exclusive-OR output. An AND logic module has a plurality of inputs and an AND output, wherein the exclusive-OR output is electrically connected to one of the plurality of inputs of the AND logic module, and the AND output provides a signal that indicates whether the first data equals the sum of the second data and third data.

Abstract: An improved arithmetic circuit for accumulative operation for use in digital signal processors, microprocessors and so forth is described, in which the pipelined control becomes effective during accumulative operation by eliminating idling stages in the pipeline structure. In accordance with the improved arithmetic circuit, during accumulative operation, the next operation is initiated with intermediate results of the current operation while the current operation is being executed and not yet completed so that it is possible to improve the speed of accumulative operation and reduce the scale of integration.

Abstract: A carry ripple adder contains five first inputs for accepting five input bits having equal significance w that are to be summed and two second inputs for accepting two carry bits having the significance w. It also contains an output for a sum bit having the significance w and two outputs for two carry bits having the significances 2w and 4w.

Abstract: An integrated circuit includes an adder having a first adder circuit for receiving a portion of the operands to be summed, along with corresponding carry-in inputs. The first adder circuit provides a sum output and carry-out outputs. A second adder circuit receives another portion of the operands to be summed, along with corresponding carry-in inputs. Multiplexers between the first and second adder circuits determine whether the carry-in inputs to the second adder circuit are the same the carry-in inputs to the first adder circuit or whether the carry-in inputs to the second adder circuit are independent.

Abstract: Circuitry for adding a first binary number (A) having a plurality of bits (a0, a1, . . . ) to a second binary number (B) having a plurality of bits (b0, b1, . . . ) to produce a third binary number (A+B) having a plurality of bits (s0, s1, . . . ) and/or a fourth binary number (A+B+1) having a plurality of bits (s0′, s1′, . . . ) and corresponding to the addition of the third binary number and one.

Abstract: Shift of input data without split by a shifter, generation of code extension data by a code extension data generator, and generation of a mask signal by a mask signal generator are carried out in parallel. The mask signal generator generates the mask signal based on an amount of shift and split mode information. An output selector replaces data shifted by the shifter with code extension data bit by bit based on the mask signal, and outputs data which are shifted and code-extended according to the split mode information and arithmetic/logical shift information. In a carry-select type adder/subtractor as another embodiment, if split parallel process is to be executed, both a pair of unit adders/subtractors execute an arithmetic operation to be carried out for the case where no carry is supplied from lower digits, and then a selector selects an arithmetic result, which is obtained when no carry is supplied from lower digits, regardless of the carry from the lower digits.

Abstract: A high-speed linear address generator (LAGEN) and method for generating a linear address are disclosed, which generator is operable to generate a linear address very quickly. In a preferred embodiment, the LAGEN has a parallel design, rather than a serial design, which allows the LAGEN to generate a linear address substantially faster than 1 nanosecond after receiving input operands. The LAGEN generates a linear address within a single clock cycle of a clock operating at 1 gigahertz (GHz). The LAGEN receives three 32-bit operands IMM[31:0], SRC1[31:0], and SRC2[31:0], and compresses them into two 32-bit operands. The LAGEN then sums the two operands producing a 32-bit result res[32:0]. The LAGEN allows for both 32-bit mode operation and 16-bit mode operation. In either mode of operation the lower 16 bits of the result, res[15:0], are output for the lower 16 bits of the generated linear address.

Abstract: A programmable multi-mode accelerator is disclosed for use with a programmable processor or microprocessor. The programmable multi-mode accelerator allows a programmable processor to execute specific algorithms, such as certain types of finite impulse response (FIR), correlation and Viterbi computations, that require low-precision operations at an extremely high rate. The accelerator extends the digital signal processor's performance into the required range for low-precision computations. The accelerator can be coupled with the main data path of a programmable processor or microprocessor and can directly read and write to the main register files of the programmable processor. In an illustrative implementation, the accelerator data path accesses its input values (source operands) directly from a main register file of the programmable processor and writes results back into a second main register file.

Abstract: A summing circuit includes a summing section and a shift bit searching section. The summing section receives an input data and a previous summation resultant data, bit-shifts the input data and the previous summation resultant data in response to first and second bit shift control signals, respectively. Also, the summing section adds the bit-shifted input data and the bit-shifted previous summation resultant data to generate a current summation resultant data and to output a part of the current summation resultant data as a shift bit calculation data. The shift bit searching section outputs the first and second bit shift control signals for addition of a next input data and the current summation resultant data to the summing section based on the shift bit calculation data.