Abstract: A matrix multiply compute apparatus and method of operation therefor. The apparatus includes a memory configured to store weight matrix elements in a first format including columns of scale factors and mantissa blocks. A first register is configured to receive the scale factors of each weight matrix column via a crossbar coupled to the memory, the converter is configured to determine max exponents using the scale factors of the columns, and a second register is configured to store the max exponents. Also, the first register is configured to receive the mantissa blocks of the columns, and the converter is configured to determine converted blocks using all of the scale factors and the mantissa blocks. A weight buffer receives the converted blocks and max exponents as the elements in a second format that is used by a compute device to determine matrix multiply outputs, which are stored in an output buffer.

Abstract: A method and device for data conversion in a matrix compute apparatus. The apparatus includes an input buffer (IB) that receives one or more matrix inputs characterized by a first format and a compute device coupled to the IB device that is configured to determine a combined matrix output. The compute device determines a first matrix output using a first input portion of the matrix input and determines a second matrix output using a second input portion. The compute device then determines a combined matrix output in a second format using the first and second matrix outputs. Within the compute device, an alignment device can determine a rounded matrix output from the combined matrix output, and a partial products reduction (PPR) device can determine a reduced matrix output in a third format using the rounded matrix output, which is stored in an output buffer (OB) coupled to the compute device.

Abstract: A communication device includes a convolutional interleaver and an encoder. The convolutional interleaver is configured to receive blocks of data defining symbol blocks that are encoded using a block code to correct an error in a block of data and to interleave the symbol blocks into a stream of interleaved symbol blocks. The encoder is configured to encode a set of symbol blocks among the interleaved symbol blocks with an error-correcting code to correct single bit errors in the set of symbol blocks. The error-correcting code is configured to generate an error-correcting block and to add the error-correcting block to the set of interleaved symbol blocks.

Abstract: A communication device includes interleaver circuitry that receives, from a host device, a first encoded data stream comprised of a plurality of symbols encoded with a first type of error correction code and interleaves the plurality of symbols of the first encoded data stream into symbol sections each including a predetermined number of symbols encoded with the first type of error correction code. Encoder circuitry encodes the first encoded data stream in accordance with a second type of error correction code different from the first type of error correction code by generating, for each of the symbol sections, an error code block corresponding to the symbols in the symbol section and outputs a second encoded data stream including the first encoded data stream and the error code block.

Abstract: The present invention is directed to communication systems and methods. According to a specific embodiment, FEC data streams from multiple FEC data lanes are received. First stage interleaving and inner encoding are performed on the FEC data streams to generate inner encoded data streams. A second stage interleaving process is performed to interleave the inner encoded data streams. There are other embodiments as well.

Abstract: The present invention is directed to lidar systems and methods thereof. More specifically, a lidar receiver converts received light signal to electrical signal. The electrical signal is converted to digital signal. Fast Fourier transform is performed on the digital signal to generate n channels of data. Constant false alarm rate detection is performed to generate n data sets, which is grouped into m clusters of data. Maximum likelihood detection is performed on the m clusters of data.

Abstract: A communication device includes interleaver circuitry that receives, from a host device, a first encoded data stream comprised of a plurality of symbols encoded with a first type of error correction code and interleaves the plurality of symbols of the first encoded data stream into symbol sections each including a predetermined number of symbols encoded with the first type of error correction code. Encoder circuitry encodes the first encoded data stream in accordance with a second type of error correction code different from the first type of error correction code by generating, for each of the symbol sections, an error code block corresponding to the symbols in the symbol section and outputs a second encoded data stream including the first encoded data stream and the error code block.

Abstract: The present invention is directed to communication systems and methods. According to a specific embodiment, FEC data streams from multiple FEC data lanes are received. First stage interleaving and inner encoding are performed on the FEC data streams to generate inner encoded data streams. A second stage interleaving process is performed to interleave the inner encoded data streams. There are other embodiments as well.

Abstract: The present invention is directed to communication systems. According to a specific embodiment, the present invention provides a network device that collects telemetry measurements related to the quality of data communication. A machine learning algorithm generates probability determinations using the telemetry measurements and an inference model. The probability determinations are used to generate alarms signals or activate repair algorithms. There are other embodiments as well.

Abstract: A communication device includes a first alignment circuit configured to receive, from a host device, a first encoded data stream including a plurality of symbols encoded with a first type of error correction code. The first alignment circuit is configured to output an aligned first encoded data stream that is aligned to boundaries between the plurality of symbols encoded with the first type of error correction code. An interleaver is configured to interleave the plurality of symbols of the aligned first encoded data stream into symbol sections each including a predetermined number of symbols encoded with the first type of error correction code. An encoder is configured to generate, for each of the symbol sections, a parity block corresponding to the symbols in the symbol section and to output a second encoded data stream including the aligned first encoded data stream and the parity block.

Abstract: Embodiments address optimization of an electrical interface between an optical host device and an optical module device at installation time. Certain methods try each entry in a set of Finite Impulse Response (FIR) filter settings at the host transmitter, while asking the module to measure the signal integrity for each. The module will then provide an indication of which entry was the best choice for signal integrity in the current hardware configuration. Note that for the module to host electrical interface, this same technique can be used in reverse, whereby the host asks the module to configure its transmitting FIR filter, and the host records and keeps track of which filter setting is the best, and then configures the module with that filter setting. In both cases, for modules supporting CMIS (Common Management Interface Specification) for module configuration and control, methods are provided.

Abstract: The present invention is directed to data communication and encoding techniques. More specifically, an embodiment of the present invention provides a communication device that aligns a data stream with RS symbols. An interleaver interleaves RS symbols to generate an interleaved RS symbol data stream. Hamming parity blocks are generated for corresponding groups of RS symbols and inserted into the interleaved RS symbol data stream. There are other embodiments as well.

Abstract: The present invention is directed to communication methods and systems thereof. In a specific embodiment, a noise cancelation system includes a slicer that processes a data stream generates both PAM symbols and error data. An RIN estimator generates RIN data based on the PAM symbols and the error data. A filter removes non-RIN information from the RIN data. The filtered RIN data includes an offset term and a gain term, which are used to remove RIN noise from the data stream. There are other embodiments as well.

Abstract: The present invention is directed to communication methods and systems thereof. In a specific embodiment, a noise cancelation system includes a slicer that processes a data stream generates both PAM symbols and error data. An RIN estimator generates RIN data based on the PAM symbols and the error data. A filter removes non-RIN information from the RIN data. The filtered RIN data includes an offset term and a gain term, which are used to remove RIN noise from the data stream. There are other embodiments as well.

Abstract: The present invention is directed to a communication signal tracking system comprising an optical receiver including one or more delay line interferometers (DLIs) configured to demultiplex incoming optical signals and a transimpedance amplifier configured to convert the incoming optical signals to incoming electrical signals. The communication signal tracking system further includes a control module configured to calculate a bit-error-rate (BER) of the incoming electrical signals before forward-error correction decoding, and use the BER as a parameter for optimizing settings of the one or more DLIs in one or more iterations in a control loop and generating a back-channel data.

Abstract: The present invention is directed to communication methods and systems thereof. In a specific embodiment, a noise cancelation system includes a slicer that processes a data stream generates both PAM symbols and error data. An RIN estimator generates RIN data based on the PAM symbols and the error data. A filter removes non-RIN information from the RIN data. The filtered RIN data includes an offset term and a gain term, which are used to remove RIN noise from the data stream. There are other embodiments as well.

Abstract: The present invention is directed to a communication signal tracking system comprising an optical receiver including one or more delay line interferometers (DLIs) configured to demultiplex incoming optical signals and a transimpedance amplifier configured to convert the incoming optical signals to incoming electrical signals. The communication signal tracking system further includes a control module configured to calculate a bit-error-rate (BER) of the incoming electrical signals before forward-error correction decoding, and use the BER as a parameter for optimizing settings of the one or more DLIs in one or more iterations in a control loop and generating a back-channel data.

Abstract: Optical fiber data communications are described. An error correction circuit can receive a signal and correct bit errors of that signal. The circuit can then determine characteristics of the signal (e.g., its bit error rate (BER)) and adjust the operations performed to correct the bit errors of the signal based on the characteristics.

Abstract: The present invention is directed to a communication signal tracking system comprising an optical receiver including one or more delay line interferometers (DLIs) configured to demultiplex incoming optical signals and a transimpedance amplifier configured to convert the incoming optical signals to incoming electrical signals. The communication signal tracking system further includes a control module configured to calculate a bit-error-rate (BER) of the incoming electrical signals before forward-error correction decoding, and use the BER as a parameter for optimizing settings of the one or more DLIs in one or more iterations in a control loop and generating a back-channel data.

Abstract: The present invention is directed to a communication signal tracking system comprising an optical receiver including one or more delay line interferometers (DLIs) configured to demultiplex incoming optical signals and a transimpedance amplifier configured to convert the incoming optical signals to incoming electrical signals. The communication signal tracking system further includes a control module configured to calculate a bit-error-rate (BER) of the incoming electrical signals before forward-error correction decoding, and use the BER as a parameter for optimizing settings of the one or more DLIs in one or more iterations in a control loop and generating a back-channel data.