A method for performing an in-memory computation includes: storing data in memory cells of a memory array, the data including weights for computation; determining whether an update command to change at least one of the weights is received; in response to receiving the update command, performing a write operation on the memory array to update the at least one weight; and disabling the write operation on the memory array until receiving a next update command to change the at least one weight.

An integrated circuit device having insulated gate field effect transistors (IGFETs) having a plurality of horizontally disposed channels that can be vertically aligned above a substrate with each channel being surrounded by a gate structure has been disclosed. The integrated circuit device may include a temperature sensor circuit and core circuitry. The temperature senor circuit may include at least one portion formed in a region other than the region that the IGFETs are formed as well as at least another portion formed in the region that the IGFETs having a plurality of horizontally disposed channels that can be vertically aligned above a substrate with each channel being surrounded by a gate structure are formed. By forming a portion of the temperature sensor circuit in regions below the IGFETs, an older process technology may be used and device size may be decreased and cost may be reduced.

An integrated circuit device having insulated gate field effect transistors (IGFETs) having a plurality of horizontally disposed channels that can be vertically aligned above a substrate with each channel being surrounded by a gate structure has been disclosed. The integrated circuit device may include a temperature sensor circuit and core circuitry. The temperature senor circuit may include at least one portion formed in a region other than the region that the IGFETs are formed as well as at least another portion formed in the region that the IGFETs having a plurality of horizontally disposed channels that can be vertically aligned above a substrate with each channel being surrounded by a gate structure are formed. By forming a portion of the temperature sensor circuit in regions below the IGFETs, an older process technology may be used and device size may be decreased and cost may be reduced.

A memory array includes memory cells forming a data word location accessed in response to a word line signal. A data sensing circuit configured to sense data on bit lines associated with the memory cells. The sensed data corresponds to a current data word stored at the data word location. A data latching circuit latches the sensed data for the current data word from the data sensing circuit. A data modification circuit then performs a mathematical modify operation on the current data word to generate a modified data word. The modified data word is then applied by a data writing circuit to the bit lines for writing back to the memory cells of the memory array at the data word location. The operations are advantageously performed within a single clock cycle.

A memory array includes memory cells forming a data word location accessed in response to a word line signal. A data sensing circuit configured to sense data on bit lines associated with the memory cells. The sensed data corresponds to a current data word stored at the data word location. A data latching circuit latches the sensed data for the current data word from the data sensing circuit. A data modification circuit then performs a mathematical modify operation on the current data word to generate a modified data word. The modified data word is then applied by a data writing circuit to the bit lines for writing back to the memory cells of the memory array at the data word location. The operations are advantageously performed within a single clock cycle.

A memory circuit includes an array of memory cells arranged with first word lines connected to a first sub-array storing less significant bits of data and second word lines connected to a second sub-array storing more significant bits of data. A row decoder circuit coupled to the first and second word lines generates word line signals. A word line gating circuit is configured to selectively gate passage of the word line signals to the second word lines for the second sub-array in response to assertion of a maximum value signal. A data modification circuit performs a mathematical operation on data read from the array of memory cells, and asserts the maximum value signal if the mathematical operation performed on the less significant bits of data from the first sub-array produces a maximum data value.

Digital compute-in-memory (DCIM) bit cell circuit layouts and DCIM array circuits for multiple operations per column are disclosed. A DCIM bit cell array circuit including DCIM bit cell circuits comprising exemplary DCIM bit cell circuit layouts disposed in columns is configured to evaluate the results of multiple multiply operations per clock cycle. The DCIM bit cell circuits in the DCIM bit cell circuit layouts each couples to one of a plurality of column output lines in a column. In this regard, in each cycle of a system clock, each of the plurality of column output lines receives a result of a multiply operation of a DCIM bit cell circuit coupled to the column output line. The DCIM bit cell array circuit includes digital sense amplifiers coupled to each of the plurality of column output lines to reliably evaluate a result of a plurality of multiply operations per cycle.

Digital compute-in-memory (DCIM) bit cell circuit layouts and DCIM array circuits for multiple operations per column are disclosed. A DCIM bit cell array circuit including DCIM bit cell circuits comprising exemplary DCIM bit cell circuit layouts disposed in columns is configured to evaluate the results of multiple multiply operations per clock cycle. The DCIM bit cell circuits in the DCIM bit cell circuit layouts each couples to one of a plurality of column output lines in a column. In this regard, in each cycle of a system clock, each of the plurality of column output lines receives a result of a multiply operation of a DCIM bit cell circuit coupled to the column output line. The DCIM bit cell array circuit includes digital sense amplifiers coupled to each of the plurality of column output lines to reliably evaluate a result of a plurality of multiply operations per cycle.

A memory architecture and a processing unit that incorporates the memory architecture and a systolic array. The memory architecture includes: memory array(s) with multi-port (MP) memory cells; first wordlines connected to the cells in each row; and, depending upon the embodiment, second wordlines connected to diagonals of cells or diagonals of sets of cells. Data from a data input matrix is written to the memory cells during first port write operations using the first wordlines and read out from the memory cells during second port read operations using the second wordlines. Due to the diagonal orientation of the second wordlines and due to additional features (e.g., additional rows of memory cells that store static zero data values or read data mask generators that generate read data masks), data read from the memory architecture and input directly into a systolic array is in the proper order, as specified by a data setup matrix.

A memory architecture and a processing unit that incorporates the memory architecture and a systolic array. The memory architecture includes: memory array(s) with multi-port (MP) memory cells; first wordlines connected to the cells in each row; and, depending upon the embodiment, second wordlines connected to diagonals of cells or diagonals of sets of cells. Data from a data input matrix is written to the memory cells during first port write operations using the first wordlines and read out from the memory cells during second port read operations using the second wordlines. Due to the diagonal orientation of the second wordlines and due to additional features (e.g., additional rows of memory cells that store static zero data values or read data mask generators that generate read data masks), data read from the memory architecture and input directly into a systolic array is in the proper order, as specified by a data setup matrix.