An IC structure includes a first SRAM cell and a second SRAM, where a layout of the second SRAM cell is a mirror image of that of the first SRAM cell about a vertical cell boundary therebetween. The first SRAM cell includes a first PD device and a second PD device disposed over a first fin and a second fin, respectively, where a portion of the first fin and a portion of the second fin corresponding to a channel region of the first and the second PD devices, respectively, each include a first stack of semiconductor layers defined by a channel width W1, a portion of the first fin and a portion of the second fin providing a source terminal of the first and the second PD devices, respectively, are each defined by a width W1′ that is enlarged with respect to the channel width W1.

An IC structure includes a first SRAM cell and a second SRAM, where a layout of the second SRAM cell is a mirror image of that of the first SRAM cell about a vertical cell boundary therebetween. The first SRAM cell includes a first PD device and a second PD device disposed over a first fin and a second fin, respectively, where a portion of the first fin and a portion of the second fin corresponding to a channel region of the first and the second PD devices, respectively, each include a first stack of semiconductor layers defined by a channel width W1, a portion of the first fin and a portion of the second fin providing a source terminal of the first and the second PD devices, respectively, are each defined by a width W1′ that is enlarged with respect to the channel width W1.

A CAM array of compare memory cell circuits includes a decode column corresponding to each set, and each set includes at least one row of the compare memory cell circuits. Each decode column receives a set clock signal addressing the corresponding set and generates a set match signal in each row of the corresponding set. A column compare circuit generates compare data indicating a bit of a compare tag. A row match circuit generates, for each row, in response to the set match signal, a row match signal indicating the compare tag matches the binary tag stored in the row. Circuits and loads in a decode column employed to generate the set clock signal correspond to circuits generating the row match signal in each column of the CAM array to reduce a timing margin of the match indication and decrease the access time for the CAM array.

A CAM array of compare memory cell circuits includes a decode column corresponding to each set, and each set includes at least one row of the compare memory cell circuits. Each decode column receives a set clock signal addressing the corresponding set and generates a set match signal in each row of the corresponding set. A column compare circuit generates compare data indicating a bit of a compare tag. A row match circuit generates, for each row, in response to the set match signal, a row match signal indicating the compare tag matches the binary tag stored in the row. Circuits and loads in a decode column employed to generate the set clock signal correspond to circuits generating the row match signal in each column of the CAM array to reduce a timing margin of the match indication and decrease the access time for the CAM array.

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.

The present disclosure relates to a method of forming a sense amplifier and a layout structure of a sense amplifier. The method includes: providing a first active region pattern layer, the first active region pattern layer includes a bridge pattern, and a first active region pattern region and a second active region pattern region; the first active region pattern region includes a first active region pattern for defining a first pull-down transistor of a first memory cell structure; the second active region pattern region includes a first symmetrical active region pattern for defining a second pull-down transistor of a second memory cell structure; and the first active region pattern and the first symmetrical active region pattern are adjacent to each other and connected through the bridge pattern, a source of the first pull-down transistor and a source of the second pull-down transistor are electrically connected through the bridge pattern.

The present disclosure relates to a method of forming a sense amplifier and a layout structure of a sense amplifier. The method includes: providing a first active region pattern layer, the first active region pattern layer includes a bridge pattern, and a first active region pattern region and a second active region pattern region; the first active region pattern region includes a first active region pattern for defining a first pull-down transistor of a first memory cell structure; the second active region pattern region includes a first symmetrical active region pattern for defining a second pull-down transistor of a second memory cell structure; and the first active region pattern and the first symmetrical active region pattern are adjacent to each other and connected through the bridge pattern, a source of the first pull-down transistor and a source of the second pull-down transistor are electrically connected through the bridge pattern.

A near memory system is provided for the calculation of a layer in a machine learning application. The near memory system includes an array of memory cells for storing an array of filter weights. A multiply-and-accumulate circuit couples to columns of the array to form the calculation of the layer.

A near memory system is provided for the calculation of a layer in a machine learning application. The near memory system includes an array of memory cells for storing an array of filter weights. A multiply-and-accumulate circuit couples to columns of the array to form the calculation of the layer.

A biased-transistor-module comprising: a module-input-terminal; a module-output-terminal; a reference-terminal; a module-supply-terminal configured to receive a supply voltage; a module-reference-voltage-terminal configured to receive a module reference voltage; a main-transistor having a main-control-terminal, a main-first-conduction-channel-terminal and a main-second-conduction-channel-terminal, wherein the main-first-conduction-channel-terminal is connected to the module-output-terminal, and the main-second-conduction-channel-terminal is connected to the reference-terminal, and the main-control-terminal is connected to an input-signal-node, wherein the input-signal-node is connected to the module-input-terminal; and a bias-circuit. The bias-circuit comprises: a first-bias-transistor; a first-bias-resistor; a second-bias-transistor; and a second-bias-resistor.