Abstract: A memory device that includes a memory array and a pre-charge selecting circuit is introduced. The memory array includes a plurality of memory cells that are coupled to a plurality of bit lines and a plurality of word lines, wherein the plurality of word lines are configured to receive an input vector. The pre-charge selecting circuit is configured to selectively pre-charge a selected bit line according to a value of the input vector. The pre-charge selecting circuit is configured to determine whether the value of the input vector is less than a predefined threshold, and generate a gated pre-charge signal to skip pre-charging the selected bit line in response to determining that the value of the input vector is less than the predefined threshold.

Abstract: A circuit includes first and second data lines, a sense amplifier including first and second input terminals, a first p-type metal-oxide-semiconductor (PMOS) transistor coupled in series with a first capacitive device between the first data line and the second input terminal, a second PMOS transistor coupled in series with a second capacitive device between the second data line and the first input terminal, a third PMOS transistor coupled between the first data line and the first input terminal, a fourth PMOS transistor coupled between the second data line and the second input terminal, a first n-type metal-oxide-semiconductor (NMOS) transistor configured to selectively couple each of the first PMOS transistor and the first capacitive device to a ground node, and a second NMOS transistor configured to selectively couple each of the second PMOS transistor and the second capacitive device to the ground node.

Abstract: An Input/Output (I/O) circuit for a memory device is provided. The I/O circuit includes a charge integration circuit coupled to a bitline of the memory device. The charge integration circuit provides a sensing voltage based on a decrease of a voltage on the bitline. A comparator is coupled to the charge integration circuit. The comparator compares the sensing voltage with a reference voltage and provides an output voltage based on the comparison. A time-to-digital converter coupled to the comparator. The time to digital convertor converts a time associated with the output voltage to a digital value.

Abstract: A memory device includes a current source and a memory array. The current source is configured to provide a current to a first node. The memory array is coupled to the current source at the first node. The memory array includes memory cells. First terminals of the memory cells are coupled to the first node. Each of the memory cells has a first resistance in response to having a first data value, and has a second resistance in response to having a second data value. The second data value is N times the first data value. The second resistance is approximately one-Nth of the first resistance, for N being a positive integer larger than one. A method of operating a memory device is also disclosed herein.

Abstract: A hybrid structure for computing-in-memory applications includes a memory cell and a digital-analog-hybrid local computing cell. The memory cell stores a weight. The digital-analog-hybrid local computing cell has a plurality of input lines, a digital output line and an analog output line. The input lines are configured to transmit a plurality of multi-bit input values. The digital-analog-hybrid local computing cell includes a digital local computing cell and a voltage local computing cell. The digital local computing cell receives the weight and is configured to generate a digital output value on the digital output line according to a higher bit of the multi-bit input values multiplied by the weight. The voltage local computing cell receives the weight and is configured to generate an analog output value on the analog output line according to a lower bit of the multi-bit input values multiplied by the weight.

Abstract: A memory includes a memory device, a reading device and a feedback device. The memory device stores a plurality of bits. The reading device includes first and second reading circuits coupled to the memory device. The second reading circuit is coupled to the first reading circuit at a first node. The first and second reading circuits cooperates with each other to generate a first voltage signal at the first node based on at least one first bit of the plurality of bits. The feedback device adjusts at least one of the first reading circuit or the second reading circuit based on the first voltage signal. The first and second reading circuits generate a second voltage signal, different from the first voltage signal, corresponding to the bits, after the at least one of the first reading circuit or the second reading circuit is adjusted by the feedback device.

Abstract: A method of storing an input data of a data set into a memory storage having bit cells. The method includes determining a bit value of a characterization bit in the input data. The method also includes writing each of remaining bits in the input data into one of the bit cells as a first state if the characterization bit has a first value, and writing each of remaining bits in the input data into the bit cells as a second state if the characterization bit has a second value that is complement to the first value. In the method, either reading the bit cell with the first state consumes less energy than reading the bit cell with the second state or the bit cell with the first state has less retention errors than the bit cell with the second state.

Abstract: A system is provided. The system includes a multiply-and-accumulate circuit and a local generator. The multiply-and-accumulate circuit is coupled to a memory array and generates a multiply-and-accumulate signal indicating a computational output of the memory array. The local generator is coupled to the memory array and generates at least one reference signal at a node in response to one of a plurality of global signals that are generated according to a number of the computational output. The local generator is further configured to generate an output signal according to the signal and a summation of the at least one reference signal at the node.

Abstract: A memory unit includes at least one memory cell and a computational cell. The at least one memory cell stores a weight. The at least one memory cell is controlled by a first word line and includes a local bit line transmitting the weight. The computational cell is connected to the at least one memory cell and receiving the weight via the local bit line. Each of an input bit line and an input bit line bar transmits a multi-bit input value. The computational cell is controlled by a second word line and an enable signal to generate a multi-bit output value on each of an output bit line and an output bit line bar according to the multi-bit input value multiplied by the weight. The computational cell is controlled by a first switching signal and a second switching signal for charge sharing.

Abstract: A circuit includes a sense amplifier, a first clamping circuit, a second clamping circuit, and a feedback circuit. The first clamping circuit includes first clamping branches coupled in parallel between the sense amplifier and a memory array. The second clamping circuit includes second clamping branches coupled in parallel between the sense amplifier and a reference array. The feedback circuit is configured to selectively enable or disable one or more of the first clamping branches or one or more of the second clamping branches in response to an output data outputted by the sense amplifier.

Abstract: A memory includes a memory device, a reading device and a feedback device. The memory device stores a plurality of bits. The reading device includes first and second reading circuits coupled to the memory device. The second reading circuit is coupled to the first reading circuit at a first node. The first and second reading circuits cooperates with each other to generate a first voltage signal at the first node based on at least one first bit of the plurality of bits. The feedback device adjusts at least one of the first reading circuit or the second reading circuit based on the first voltage signal. The first and second reading circuits generate a second voltage signal, different from the first voltage signal, corresponding to the bits, after the at least one of the first reading circuit or the second reading circuit is adjusted by the feedback device.

Abstract: The disclosure introduces a shift register is configured to enter a low power mode by disabling a portion of flip-flops (FFs) that handles upper bits of input data. The shift register includes first FF(s), second FF(s) and gating circuit. The first flip-flop (FF), includes input terminal coupled to first portion of input data. The second FF includes input terminal coupled to second portion of input data, an output terminal, a clock terminal coupled to a clock signal, a power terminal coupled to a supply power. The second portion of the input data is subsequent to the first portion of the input data. The gating circuit is coupled to the output terminal of the first FF, and configured to disable the second FF for storing the second portion of a subsequent input data according to output data currently being stored in the first FF.

Abstract: A memory device includes a memory array, a multiply-accumulate (MAC) circuit and an encoder-decoder circuit. The MAC circuit performs a MAC operation on an encoded weight data stored in the memory array and an input data to generate a partial MAC result. An encoder of the encoder-decoder circuit is configured to encode m weight bits among n weight bits of weight data according to an encryption key to generate the encoded weight data, wherein m and n are positive integers, and m is less than n. A decoder of the encoder-decoder circuit is configured to detect an error in the partial MAC result according to the encryption key to generate a decoded partial MAC result.

Abstract: A computing circuit is provided. The computing circuit is disposed in a memory device and electrically coupled to a memory cell of the memory device. The computing circuit includes a weight decoder, a multiplier, an adder tree, and an accumulator. The weight decoder is configured to obtain a compressed weight from the memory cell and generate a decoded weight based on the compressed weight. The multiplier is configured to generate a partial-product by multiplying an input signal with the decoded weight. The adder tree is configured to generate a partial-sum by performing an addition operation based on the partial-product. The accumulator is configured to generate an accumulated sum by performing an accumulation operation based on the partial-sum and output an output signal based on the accumulated sum. The accumulated sum is left shifted based on a shift signal.

Abstract: A floating point pre-alignment structure for computing-in-memory applications includes a time domain exponent computing block and an input mantissa pre-align block. The time domain exponent computing block is configured to compute a plurality of original input exponents and a plurality of original weight exponents to generate a plurality of flags. Each of the flags is determined by adding one of the original input exponents and one of the original weight exponents. The input mantissa pre-align block is configured to receive a plurality of original input mantissas and shift the original input mantissas according to the flags to generate a plurality of weighted input mantissas, and sparsity of the weighted input mantissas is greater than sparsity of the original input mantissas. Each of the flags has a negative correlation with a sum of the one of the original input exponents and the one of the original weight exponents.

Abstract: A memory unit with time domain edge delay accumulation for computing-in-memory applications is controlled by a first word line and a second word line. The memory unit includes at least one memory cell, at least one edge-delay cell multiplexor and at least one edge-delay cell. The at least one edge-delay cell includes a weight reader and a driver. The weight reader is configured to receive a weight and a multi-bit analog input voltage and generate a multi-bit voltage according to the weight and the multi-bit analog input voltage. The driver is connected to the weight reader and configured to receive an edge-input signal. The driver is configured to generate an edge-output signal having a delay time according to the edge-input signal and the multi-bit voltage. The delay time of the edge-output signal is positively correlated with the multi-bit analog input voltage multiplied by the weight.

Abstract: An Input/Output (I/O) circuit for a memory device is provided. The I/O circuit includes a charge integration circuit coupled to a bitline of the memory device. The charge integration circuit provides a sensing voltage based on a decrease of a voltage on the bitline. A comparator is coupled to the charge integration circuit. The comparator compares the sensing voltage with a reference voltage and provides an output voltage based on the comparison. A time-to-digital converter coupled to the comparator. The time-to-digital convertor converts a time associated with the output voltage to a digital value.

Abstract: A circuit includes a reference voltage node, first and second data lines, a sense amplifier, first and second switching devices coupled between the first and second data lines and first and second input terminals of the sense amplifier, third and fourth switching devices coupled between the first and second data lined and first and second nodes, fifth and sixth switching devices coupled between the first and second nodes and the reference voltage node, and first and second capacitive devices coupled between the first and second nodes and second and first input terminals. Each of the first through fourth switching devices is switched on and each of the fifth and sixth switching devices is switched off in a first operational mode, and each of the first through fourth switching devices is switched off and each of the fifth and sixth switching devices is switched on in a second operational mode.

Abstract: A computation apparatus and a computation method with input swapping are provided. The computation apparatus includes a non-zero detection circuit, a swapper policy circuit, a swapper matrix circuit, and an adder tree. The non-zero detection circuit is configured to receive input vectors, inspect non-zero operands in the input vectors and generate a non-zero indicative signal indicating the non-zero operands. The swapper policy circuit is configured to receive and interpret the non-zero indicative signal, and generate multiplexer (MUX) selection signals for swapping the non-zero operands according to a set of swapping policies. The swapper matrix circuit is configured to receive the input vectors and the MUX selection signal, and perform swapping on operands in the input vectors according to the MUX selection signal. The adder tree is configured to receive the input vectors with the swapped operands and perform additions on the input vectors to output a computation result.

Abstract: A memory device and an operation method thereof are provided. The memory device includes memory cells, each having a static random access memory (SRAM) cell and a non-volatile memory cell. The SRAM cell is configured to store complementary data at first and second storage nodes. The non-volatile memory cell is configured to replicate and retain the complementary data before the SRAM cell loses power supply, and to rewrite the replicated data to the first and second storage nodes of the SRAM cell after the power supply of the SRAM cell is restored.