Electronics and Telecommunications Research Institute

Young-Su Kwon

Jin Ho Han

Jin-Ho Han

Nak Woong Eum

Min-Hyung Cho

Kyung Jin Byun

Young-deuk Jeon

Jeongmin Yang

Min-Seok Choi

Yong Cheol Peter Cho

Chun-Gi Lyuh

Yi-Gyeong Kim

Hyun Mi Kim

Hyun Kyu Yu

Sung Weon Kang

Joo Hyun Lee

Chan Soo Kim

Su-Jin Park

Kyoung Seon Shin

Jaehoon Chung

Jae-Woong Choi

Hyuk Joo Kim

Seong Min Kim

Bon Tae Koo

Young-Deuk Jeon

Seon-Ho Han

In San Jeon

Kyung-Jin Byun

Suk Ho Lee

Nak-Woong Eum

Bon-Tae Koo

Woojoo Lee

Joo-Hyun Lee

Young-Seok Baek

Yong-Cheol Cho

Disclosed herein is an apparatus for receiving data from memory. The apparatus receives a data signal and a clock signal output from memory and includes a Decision Feedback Equalizer (DFE) including two or more differential signal path units configured to determine and output an output value corresponding to the data signal. Each of the two or more differential signal path units may determine a current output value by reflecting a previous output value fed back from a different one of the two or more differential signal path units in such a way that they operate at different clocks, and may include an offset control unit configured to adjust an offset at an input stage and a feedback control unit configured to change a load of an output stage using the previous output value fed back from the different one of the two or more differential signal path units.

Apparatus for receiving data from memory

Disclosed herein are a method for supporting cache coherency based on virtual addresses for an artificial intelligence processor having large on-chip memory and an apparatus for the same. The method for supporting cache coherency according to an embodiment of the present disclosure includes, by an artificial intelligence processor including multiple processor cores and multiple caches, setting external memory address areas which do not overlap each other for respective multiple caches; and providing virtual addresses with which the multiple processor cores access the multiple caches.

Method for supporting cache coherency based on virtual addresses for artificial intelligence processor having large on-chip memory and apparatus for the same

An apparatus for receiving a strobe signal may include an amplifier for amplifying a strobe signal input thereto, an offset generator for controlling the setting of a threshold for detecting a preamble signal by generating an offset for the amplifier, and a preamble detector for detecting a first preamble signal occurring at a point at which the amplified strobe signal is equal to or greater than the threshold and turning off the offset generator when the first preamble signal is detected.

Apparatus and method for receiving strobe signal

A reorganizable neural network computing device is provided. The computing device includes a data processing array unit including a plurality of operators disposed at locations corresponding to a row and a column. One or more chaining paths which transfer the first input data from the operator of the first row of the data processing array to the operator of the second row are optionally formed. The plurality of first data input processors of the computing device transfer the first input data for a layer of the neural network to the operators along rows of the data processing array unit, and the plurality of second data input processors of the computing device transfer the second input data to the operators along the columns of the data processing array.

Reorganizable data processing array for neural network computing

Disclosed herein are a duty cycle monitoring method and apparatus for a memory interface, including receiving a clock signal as input and generating a first delay time offset and a second delay time offset, receiving the clock signal and the first delay time offset and then outputting a first delayed signal, receiving the first delayed signal and the second delay time offset and then outputting a second delayed signal, receiving the clock signal and the second delayed signal and then outputting a delay value corresponding to a half-period of the clock signal, and monitoring, based on the first delayed signal, whether a duty cycle of the clock signal conforms to a duty cycle specification.

Duty cycle monitoring method and apparatus for memory interface

Disclosed are a digital CDR circuit and a feedback loop circuit including the same. The digital CDR circuit includes a phase detector that receives an input signal and outputs a phase detection result signal corresponding to a determination result for a sampling time based on the input signal, a charge pump that receives the phase detection result signal and outputs an amplified signal obtained by multiplying the phase detection result signal by a gain, a loop filter that receives the amplified signal and filters the amplified signal to output a filtered signal, and a phase shift control code generator that generates a control signal for controlling a phase of a signal based on the filtered signal, and the input signal includes plural data signals and plural error signals, and the data signals and the error signals are digital signals which are quantized based on a signal magnitude.

Digital clock and data recovery circuit and feedback loop circuit including the same

Provided is a memory interface device. A memory interface device, comprising: a DQS input buffer configured to receive input data strobe signals and output a first intermediate data strobe signal, the DQS input buffer providing a static offset; an offset control circuit configured to receive the first intermediate data strobe signal and output a second intermediate data strobe signal; and a duty adjustment buffer configured to receive the second intermediate data strobe signal and output a clean data strobe signal, wherein the offset control circuit provides a dynamic offset using the clean data strobe signal.

Memory interface device

Disclosed is a parallel processor. The parallel processor includes a processing element array including a plurality of processing elements arranged in rows and columns, a row memory group including row memories corresponding to rows of the processing elements, a column memory group including column memories corresponding to columns of the processing elements, and a controller to generate a first address and a second address, to send the first address to the row memory group, and to send the second address to the column memory group. The controller supports convolution operations having mutually different forms, by changing a scheme of generating the first address.

Parallel processor, address generator of parallel processor, and electronic device including parallel processor

Disclosed herein is a heterogeneous system based on unified virtual memory. The heterogeneous system based on unified virtual memory may include a host for compiling a kernel program, which is source code of a user application, in a binary form and delivering the compiled kernel program to a heterogenous system architecture device, the heterogenous system architecture device for processing operation of the kernel program delivered from the host in parallel using two or more different types of processing elements, and unified virtual memory shared between the host and the heterogenous system architecture device.

Host apparatus, heterogeneous system architecture device, and heterogeneous system based on unified virtual memory

An electronic device includes first to n-th cells ('n' is an integer of 2 or more) that receive spatial-temporal input signals that indicate an event unit in a time window, a summation circuit that sums first to n-th cell signals recorded in the first to n-th cells for each of first to m-th unit times ('m' is an integer of 2 or more) dividing the time window to generate first to m-th summation signals, and an encoding circuit that compares each of the first to m-th summation signals with a threshold value to encode the spatial-temporal input signals into a code of the event unit.

Electronic device for encoding event indicated by spatial-temporal input signals and operating method thereof

Provided is a neural network accelerator which performs a calculation of a neural network provided with layers, the neural network accelerator including a kernel memory configured to store kernel data related to a filter, a feature map memory configured to store feature map data which are outputs of the layers, and a Processing Element (PE) array including PEs arranged along first and second directions, wherein each of the PEs performs a calculation using the feature map data transmitted in the first direction from the feature map memory and the kernel data transmitted in the second direction from the kernel memory, and transmits a calculation result to the feature map memory in a third direction opposite to the first direction.

Neural network accelerator including bidirectional processing element array

A processing element and an operating method thereof in a neural network are disclosed. The processing element may include a first multiplexer selecting one of a first value stored in a first memory and a second value stored in a second memory, a second multiplexer selecting one of a first data input signal and an output value of the first multiplexer, a third multiplexer selecting one of the output value of the first multiplexer and a second data input signal, a multiplier multiplying an output value of the second multiplexer by an output value of the third multiplexer, a fourth multiplexer for selecting one of the output value of the second multiplexer and an output value of the multiplier, and a third memory storing an output value of the fourth multiplexer.

Processing element and operating method thereof in neural network

Disclosed is a cache including a dataflow controller for transmitting first data to a first processor and receiving second data from the first processor, an external direct memory access (DMA) controller for receiving the first data from an external memory to transmit the first data to the dataflow controller and receiving the second data from the dataflow controller to transmit the second data to the external memory, a scratchpad memory for storing the first data or the second data transmitted between the dataflow controller and the external DMA controller, a compression/decompression device for compressing data to be transmitted from the scratchpad memory to the external memory and decompressing data transmitted from the external memory to the scratchpad memory, and a transfer state buffer for storing transfer state information associated with data transfer between the dataflow controller and the external DMA controller.

Cache for artificial intelligence processor

Provided is an image recognition processor. The image recognition processor includes a plurality of nano cores arranged in rows and columns and configured to perform a pattern recognition operation on an input feature using a kernel coefficient in response to each instruction, an instruction memory configured to provide the instruction to each of the plurality of nano cores, a feature memory configured to provide the input feature to each of the plurality of nano cores, a kernel memory configured to provide the kernel coefficients to the plurality of nano cores, and a functional safety processor core configured to receive a result of a pattern recognition operation outputted from the plurality of nano cores to detect the presence of a recognition error, and perform a fault tolerance function on the detected recognition error.

Image recognition processor including functional safety processor core and operation method thereof

Disclosed is a receiving circuit, which includes a hysteresis detector that receives an input signal corresponding to a first voltage level and outputs a detection signal having a first threshold voltage and a second threshold voltage, and a level shifter that receives the detection signal, converts the first voltage level of the detection signal to a second voltage level higher than the first voltage level so as to be output as an output signal, and outputs a feedback signal of the second voltage level, and the hysteresis detector receives the feedback signal from the level shifter and adjusts the first threshold voltage and the second threshold voltage based on the feedback signal.

Growing community of inventors

Daejeon, South Korea

Young-Su Kwon