Pipeline arithmetic adder and multiplier

Claim 2

A pipeline adder for developing a sum of two numbers having M bits each, where M is a number, comprising: a plurality of cells arranged in rows and columns, with each cell operating on data signal related to said two numbers to develop output data signals related to said two numbers, with the first row of cells being responsive to said two numbers, each subsequent row of cells being responsive to developed data signals emanating solely from the immediately previous row of cells, and the last row of cells developing said sum of said two numbers; N-1 least significant cells in each of said rows providing a preselected delay, N being a number, and M-N+1 next higher significance cells in each of said rows performing arithmetic addition, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance cells that perform arithmetic addition developing a binary sum signal output responsive to its inputs and a carry signal output responsive to its inputs; and the most significant cell in the third and each subsequent of said rows, responsive to an output signal of a most significant cell in the immediately previous row and the same column and to a carry signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the OR function of said cell's input signals.

Claim 3

The adder of claim 1 wherein said cells include means for reclocking output signals of said cells.

Claim 4

The adder of claim 1 wherein said cells that perform arithmetic addition are half adders.

Claim 5

The adder of claim 1 wherein cells in selected ones of said rows include means for reclocking signals of said cells.

Claim 6

The adder of claim 1 wherein said cells include means for reclocking, and cells in at least one of said rows include means for excluding said means for reclocking from the signal path of said cells.

Claim 7

The adder of claim 1 wherein said cells include means for reclocking, and cells in at least one of said rows include switch means for bypassing said means for reclocking.

Claim 8

The adder of claim 1 wherein some of said arithmetic addition cells comprise a clocked sum input, a clocked carry input, a clocked sum output, a clocked carry output, an unclocked sum output, and an unclocked carry output, with said clocked sum output developed in accordance with the exclusive OR function of said unclocked carry input with an exclusive OR function of said clocked sum input and said clocked carry input, with said clocked carry output developed in accordance with the AND function of said exclusive OR of said clocked sum input and said clocked carry input with said unclocked carry input, and with said unclocked carry output developed in accordance with the AND function of said clocked sum input and said clocked carry input.

Claim 9

The adder of claim 1 wherein cells of every other one of said rows include means for reclocking signals of said cells and cells of other ones of said rows are unclocked.

Claim 10

A pipeline adder for developing a sum of two numbers having M bits each, where M is a number, comprising: a plurality of cells arranged in rows and columns, with each cell operating on data signals related to said two numbers to develop output data signals related to said two numbers, with the first row of cells being responsive to said two numbers, each subsequent row of cells being responsive to developed data signals emanating solely from the immediately previous row of cells, and the last row of cells developing said sum of said two numbers; N-1 least significant cells in each of said rows providing a preselected delay and signal inversion, N being a number, and M-N+1 next higher significance cells, in each of said rows performing arithmetic addition, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum inverse signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry inverse signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance arithmetic addition cells developing a sum inverse signal output equal to the Exclusive NOR function of said sum input and said carry inputs, and a carry inverse signal output equal to the NAND function of said sum inputs and said at least one carry input; and the most significant cell in the third and each subsequent of said rows, responsive to an output signal of a most significant cell in the immediately previous row and the same column and to a carry inverse signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the NOR function of said cell's input signals.

Claim 11

A pipeline adder for developing a sum of two numbers having M bits each, where M is a number, comprising: a plurality of cells arranged in rows and columns, with each cell operating on data signals related to said two numbers to develop output data signals related to said two numbers, with the first row of cells being responsive to said two numbers, each subsequent row of cells being responsive to developed data signals emanating solely from the immediately previous row of cells, and the last row of cells developing said sum of said two numbers; N-1 least significant cells in each of said rows providing a preselected delay and signal inversion, N being a number, and M-N+1 next higher significance cells, in each of said rows performing arithmetic addition, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum inverse signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry inverse signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance arithmetic addition cells developing a sum inverse signal output equal to the Exclusive NOR function of said sum input and said carry inputs, and a carry inverse signal output equal to the NAND function of said sum inputs and said at least one carry input; and the most significant cell in the third and each subsequent of said rows, responsive to an output signal of a most significant cell in the immediately previous row and the same column and to a carry inverse signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the NOR function of said cell's input signals.

Claim 12

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion.

Claim 13

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion.

Claim 14

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline adder section further comprises: a plurality of cells arranged in rows and columns, with each cell operating on data signals related to said two numbers to develop output data signals related to said two numbers, with the first row of cells being responsive to said two numbers, each subsequent row of cells being responsive to developed data signals emanating solely from the immediately previous row of cells, and the last row of cells developing said sum of said two numbers; N-1 least significant cells in each of said rows providing a preselected delay, N being a number, and M-N+1 next higher significance cells in each of said rows performing arithmetic addition and signal inversion, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance arithmetic addition cells developing a sum signal output equal to the Exclusive OR function of said sum input and said carry inputs, and a carry signal output equal to the AND function of said sum inputs and said at least one carry input; and the most significant cell in the third and each subsequent of said rows responsive to an output signal of a most significant cell in the immediately previous row and the same column and to a carry signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the OR function of said cell's input signals.

Claim 15

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline adder section further comprises: a plurality of cells arranged in rows and columns, with each cell operating on data signals related to said two numbers to develop output data signals related to said two numbers, with the first row of cells being responsive to said two numbers, each subsequent row of cells being responsive to developed data signals emanating solely from the immediately previous row of cells, and the last row of cells developing said sum of said two numbers; N-1 least significant cells in each of said rows providing a preselected delay, N being a number, and M-N+1 next higher significance cells in each of said rows performing arithmetic addition and signal inversion, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance arithmetic addition cells developing a sum signal output equal to the Exclusive OR function of said sum input and said carry inputs, and a carry signal output equal to the AND function of said sum inputs and said at least one carry input; and the most significant cell in the third and each subsequent of said rows responsive to an output signal of a most significant cell in the immediately previous row and the same column and to a carry signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the OR function of said cell's input signals.

Claim 16

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline multiplier further comprises delay cells, half adder cells and full adder cells arranged in rows, with each of said delay, half adder and full adder cells including means for reclocking the output signals of said cells and with each of said delay, half adder and full adder cells in at least one of said rows including switch means responsive to a control signal for excluding said means for reclocking from the signal path of said output signals of said cells.

Claim 17

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline multiplier further comprises delay cells, half adder cells and full adder cells arranged in rows, with each of said delay, half adder and full adder cells including means for reclocking the output signals of said cells and with each of said delay, half adder and full adder cells in at least one of said rows including switch means responsive to a control signal for excluding said means for reclocking from the signal path of said output signals of said cells.

Claim 18

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline adder section and said delay section comprise rows of cells including means for reclocking the output signals of said cells, and switch means included in cells of at least one of said rows, responsive to a control signal, for bypassing said means for reclocking.

Claim 19

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline adder section and said delay section comprise rows of cells including means for reclocking the output signals of said cells, and switch means included in cells of at least one of said rows, responsive to a control signal, for bypassing said means for reclocking.

Claim 20

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline adder section further comprises: a plurality of cells arranged in rows and columns; N-1 least significant cells in each of said rows providing a preselected delay and signal inversion, and M-N+1 next higher significance cells in each of said rows performing arithmetic addition and signal inversion, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum inverse signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry inverse signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance cells that perform arithmetic addition developing a sum inverse signal output equal to the Exclusive NOR function of said sum input and said carry inputs, and a carry inverse signal output equal to the NAND function of said sum inputs and said carry input; and the most significant cell in the third and each subsequent of said rows responsive to an output signal of a most significant cell in the immediately previous row and the same column and to carry signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the NOR function of said cell's input signals.

Claim 21

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; and a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; wherein said pipeline adder section further comprises: a plurality of cells arranged in rows and columns; N-1 least significant cells in each of said rows providing a preselected delay and signal inversion, and M-N+1 next higher significance cells in each of said rows performing arithmetic addition and signal inversion, where N is the row number beginning with 1; each of said higher significance cells that perform arithmetic addition in rows other than the first row having a sum input connected to a sum inverse signal output developed by a cell in the same column and in the immediately preceding row, and having a carry input connected to a carry inverse signal output developed by a cell in a column of one bit lower significance and in the immediately preceding row, with each of said higher significance cells that perform arithmetic addition developing a sum inverse signal output equal to the Exclusive NOR function of said sum input and said carry inputs, and a carry inverse signal output equal to the NAND function of said sum inputs and said carry input; and the most significant cell in the third and each subsequent of said rows responsive to an output signal of a most significant cell in the immediately previous row and the same column and to carry signal from an arithmetic addition cell in said immediately previous row and in a column of one bit lower significance, developing a cell output signal in accordance with the NOR function of said cell's input signals.

Claim 22

A pipeline multiplier for developing a product of a multiplicand having M bits and a multiplier having N bits, M and N each being a number, comprising: a para-multiplication section of N-1 rows, all but the first of said rows being responsive to a partial product signal from the immediately previous row and to said multiplier, said partial product signals containing a plurality of sum signals and a plurality of carry signals; a delay section responsive to N-1 least significant ones of the sum signals on the partial product signal of the last of said rows, for delaying said N-1 least significant ones of said partial sum output signals; and a pipeline adder section responsive to remaining ones of said sum signals and to said carry signals the sum signals on the partial product signal of the last of said rows for combining said remaining ones of said partial sum output signals with said partial carry output signals.

Claim 23

A pipeline multiplier for developing a product of a multiplicand having M bits and a multiplier having N bits, M and N each being a number, comprising: a para-multiplication section of N-1 rows, all but the first of said rows being responsive to a partial product signal from the immediately previous row and to said multiplier, said partial product signals containing a plurality of sum signals and a plurality of carry signals; a delay section responsive to N-1 least significant ones of the sum signals on the partial product signal of the last of said rows, for delaying said N-1 least significant ones of said partial sum output signals; and a pipeline adder section responsive to remaining ones of said sum signals and to said carry signals the sum signals on the partial product signal of the last of said rows for combining said remaining ones of said partial sum output signals with said partial carry output signals.

Claim 24

The multiplier of claim 15 wherein each row on said para-multiplication section is responsive to signals developed by the immediately preceding row in said para-multiplication section and to a signal corresponding to said multiplicand ANDed with a selected bit of said multiplier.

Claim 25

The multiplier of claim 15 wherein said multiplier possesses bits x 0 , x 1 , x 2 , . . . x N-1 , said multiplicand possesses bits y 0 , y 1 , y 2 , . . . y M-1 , the first row of said para-multiplication section is responsive to a signal corresponding to said multiplicand ANDed with bit x 0 of said multiplier and to a signal corresponding to said multiplicand ANDed with bit x 1 of said multiplier, and each subsequent row j is responsive to signals developed by row j-1 and to a signal corresponding to said multiplicand ANDed with bit x j of said multiplier.

Claim 26

A pipeline multiplier for developing a product of a multiplicand having M bits and a multiplier having N bits, M and N each being a number comprising: a para-multiplication section of N-1 rows, all but the first of said rows being responsive to a partial product signal from the immediately previous row and to said multiplier, said partial product signals containing a plurality of sum signals and a plurality of carry signals; a delay section responsive to N-1 least significant ones of the sum signals on the partial product signal of the last of said rows, for delaying said N-1 least significant ones of said partial sum output signals; a pipeline adder section responsive to remaining ones of said sum signals and to said carry signals the sum signals on the partial product signal of the last of said rows for combining said remaining ones of said partial sum output signals with said partial carry output signals; wherein said multiplier possesses bit x 0 , x 1 , x 2 , . . . x N-1 , said multiplicand possesses bits y 0 , y 1 , y 2 , . . . y M-1 , the first row of said para-multiplication section is responsive to a signal corresponding to said multiplicand ANDed with bit x 0 of said multiplier and to a signal corresponding to said multiplicand ANDed with bit x 1 of said multiplier, and each subsequent row j is responsive to signals developed by row j-1 and to a signal corresponding to said multiplicand ANDed with bit x j of said multiplier; and means for inverting the ANDed results of bit y M-1 with bits x i , where i=0,1,2, . . . N-2 and for inverting the ANDed results of bits y k , where k=0,1,2, . . . M-2, with bit x N-1 , with respect to ANDed results of other bits y k with bits x i .

Claim 27

A pipeline multiplier for developing a product of a multiplicand having M bits and a multiplier having N bits, M and N each being a number comprising: a para-multiplication section of N-1 rows, all but the first of said rows being responsive to a partial product signal from the immediately previous row and to said multiplier, said partial product signals containing a plurality of sum signals and a plurality of carry signals; a delay section responsive to N-1 least significant ones of the sum signals on the partial product signal of the last of said rows, for delaying said N-1 least significant ones of said partial sum output signals; a pipeline adder section responsive to remaining ones of said sum signals and to said carry signals the sum signals on the partial product signal of the last of said rows for combining said remaining ones of said partial sum output signals with said partial carry output signals; wherein said multiplier possesses bit x 0 , x 1 , x 2 , . . . x N-1 , said multiplicand possesses bits y 0 , y 1 , y 2 , . . . y M-1 , the first row of said para-multiplication section is responsive to a signal corresponding to said multiplicand ANDed with bit x 0 of said multiplier and to a signal corresponding to said multiplicand ANDed with bit x 1 of said multiplier, and each subsequent row j is responsive to signals developed by row j-1 and to a signal corresponding to said multiplicand ANDed with bit x j of said multiplier; and means for inverting the ANDed results of bit y M-1 with bits x i , where i=0,1,2, . . . N-2 and for inverting the ANDed results of bits y k , where k=0,1,2, . . . M-2, with bit x N-1 , with respect to ANDed results of other bits y k with bits x i .

Claim 28

A pipeline multiplier for developing a product of a multiplicand having M bits and a multiplier having N bits, M and N each being a number, comprising: a para-multiplication section of N-1 rows, all but the first of said rows being responsive to a partial product signal from the immediately previous row and to said multiplier, said partial product signals containing a plurality of sum signals and a plurality of carry signals; a delay section responsive to N-1 least significant ones of the sum signals on the partial product signal of the last of said rows, for delaying said N-1 least significant ones of said partial sum output signals; and a pipeline adder section responsive to remaining ones of said sum signals and to said carry signals the sum signals on the partial product signal of the last of said rows for combining said remaining ones of said partial sum output signals with said partial carry output signals; wherein said multiplier possesses bits x 0 , x 1 , x 2 , . . . x N-1 , said multiplicand possesses bits y 0 , y 1 , y 2 , . . . y M-1 , the first row of said para-multiplication section is responsive to a signal corresponding to said multiplicand ANDed with bit x 0 of said multiplier and to a signal corresponding to said multiplicand ANDed with bit x 1 of said multiplier, and each subsequent row j is responsive to signals developed by row j-1 and to a signal corresponding to said multiplicand ANDed with bit x j of said multiplier; and wherein said delay section and said pipeline adder section include a preselected number of rows, and where each row j of said multiplier comprises cells c i , where i=0,1,2, . . . and c 0 is the least significant cell, and each cell c i of row j≠0 where j=0 represents the first row receives signals from cell c i or cell c i-1 , or both cells c i and c i-1 of row j+1, and each cell c i of row j other than the last row of said multiplier provided signals to cell c i of row j+1.

Claim 29

A pipeline multiplier for developing a product of a multiplicand having M bits and a multiplier having N bits, M and N each being a number, comprising: a para-multiplication section of N-1 rows, all but the first of said rows being responsive to a partial product signal from the immediately previous row and to said multiplier, said partial product signals containing a plurality of sum signals and a plurality of carry signals; a delay section responsive to N-1 least significant ones of the sum signals on the partial product signal of the last of said rows, for delaying said N-1 least significant ones of said partial sum output signals; and a pipeline adder section responsive to remaining ones of said sum signals and to said carry signals the sum signals on the partial product signal of the last of said rows for combining said remaining ones of said partial sum output signals with said partial carry output signals; wherein said multiplier possesses bits x 0 , x 1 , x 2 , . . . x N-1 , said multiplicand possesses bits y 0 , y 1 , y 2 , . . . y M-1 , the first row of said para-multiplication section is responsive to a signal corresponding to said multiplicand ANDed with bit x 0 of said multiplier and to a signal corresponding to said multiplicand ANDed with bit x 1 of said multiplier, and each subsequent row j is responsive to signals developed by row j-1 and to a signal corresponding to said multiplicand ANDed with bit x j of said multiplier; and wherein said delay section and said pipeline adder section include a preselected number of rows, and where each row j of said multiplier comprises cells c i , where i=0,1,2, . . . and c 0 is the least significant cell, and each cell c i of row j≠0 where j=0 represents the first row receives signals from cell c i or cell c i-1 , or both cells c i and c i-1 of row j+1, and each cell c i of row j other than the last row of said multiplier provided signals to cell c i of row j+1.

Claim 30

The multiplier of claim 19 wherein for a preselected value of j, cell c i-j of row i=j, j+1, j+2, . . . is an adder cell that provides an input port to an external signal z i-j .

Claim 31

A recursive pipeline adder for developing an accumulated sum numbers having M bits each, where M is a number comprising: a plurality of cells arranged in rows and columns, with the first row of cells being responsive to said two numbers, the next M+1 rows of cells having one cell in each column being responsive to developed data signals emanating the row immediately preceding said cell and from the output of said recursive adder at said column, and having the remaining cells in the row responsive to developed data signals emanating solely from the immediately previous row of cells.

Claim 32

A recursive pipeline adder for developing an accumulated sum numbers having M bits each, where M is a number comprising: a plurality of cells arranged in rows and columns, with the first row of cells being responsive to said two numbers, the next M+1 rows of cells having one cell in each column being responsive to developed data signals emanating the row immediately preceding said cell and from the output of said recursive adder at said column, and having the remaining cells in the row responsive to developed data signals emanating solely from the immediately previous row of cells.

Claim 33

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; and wherein said cells in said para-multiplier section in one of said rows, i, are connected to cells in a subsequent one of said rows, i+1, where i is an integer, cells of every other row of said cells are clocked, and cells of remaining one of said rows of cells are unclocked.

Claim 34

An arithmetic unit for multiplying signals having M bits each where M is a number, and developing a product signal, comprising: an M bit pipeline para-multiplier subsection developing 2M-1 sum output signals and M-1 carry output signals having a plurality of rows, and a plurality of cells within each row, with each of said cells in a row developing a sum signal corresponding to the inverse of the arithmetic sum of the cell's inputs and a carry inverse signal corresponding to the inverse of the arithmetic carry of the cell's inputs, said sum signal and said carry signal of a cell in one row being connected to cells in the same following row; a delay section having M inputs and M outputs, responsive to a clock and to M least significant sum output signals of said pipeline multiplier, and imposing a delay of M clock pulses between signals at said M inputs and corresponding signals at said M outputs; a pipeline adder section connected to M-1 most significant of said sum signals and to said M-1 carry output signals and developing the most significant portion; and wherein said cells in said para-multiplier section in one of said rows, i, are connected to cells in a subsequent one of said rows, i+1, where i is an integer, cells of every other row of said cells are clocked, and cells of remaining one of said rows of cells are unclocked.