According to an example aspect of the present invention, there is provided a method, comprising: receiving an input ordered set of transactions after a genesis block or a preceding compressed block in a chain of blocks, generating a compressed block on the basis of the input ordered set of transactions, wherein processing of the compressed block results to an equivalent final state as processing of the input ordered set of transactions, and providing the compressed block to a distributed network for establishing a new chain epoch and replacing a set of uncompressed blocks associated with the input ordered set of transactions.

A method for presenting output returned by a command-line interface is disclosed. In one embodiment, such a method submits, by way of a command-line interface (CLI), a command to retrieve a list of resources. The method further submits, in association with the command, an indicator to compress the list. In response to receiving the command and indicator, the method retrieves the list of resources and compresses the list such that resources with identical attributes are presented as a single tuple in the list. Such a tuple may, in certain embodiments, identify the resources with identical attributes as a range of resources and/or as a comma delimited list of resources. The tuple may also, in certain embodiments, identify how many resources with identical attributes are represented by the tuple. A corresponding system and computer program product are also disclosed.

Apparatus and methods for training a neural network accelerator using quantized precision data formats are disclosed, and in particular for storing activation values from a neural network in a compressed format having lossy or non-uniform mantissas for use during forward and backward propagation training of the neural network. In certain examples of the disclosed technology, a computing system includes processors, memory, and a compressor in communication with the memory. The computing system is configured to perform forward propagation for a layer of a neural network to produced first activation values in a first block floating-point format. In some examples, activation values generated by forward propagation are converted by the compressor to a second block floating-point format having a non-uniform and/or lossy mantissa. The compressed activation values are stored in the memory, where they can be retrieved for use during back propagation.

Methods, apparatus, systems and articles of manufacture are disclosed to facilitate malware detection using compressed data. An example apparatus includes an input processor to obtain a model, the model identifying a first sequence associated with a first trace of data known to be repetitive, a sequence identifier to identify a second sequence associated with a second trace of data, a comparator to compare the first sequence with the second sequence, and an output processor to when the first sequence matches the second sequence, transmit an encoded representation of the second sequence to the central processing facility using a first channel of communication, and when the first sequence fails to match the second sequence, transmit the second sequence to the central processing facility using a second channel of communication, the second sequence to be analyzed by the central processing facility to identify whether the second sequence is indicative of malware.

Described herein is a neural network accelerator with a set of neural processing units and an instruction set for execution on the neural processing units. The instruction set is a compact instruction set including various compute and data move instructions for implementing a neural network. Among the compute instructions are an instruction for performing a fused operation comprising sequential computations, one of which involves matrix multiplication, and an instruction for performing an elementwise vector operation. The instructions in the instruction set are highly configurable and can handle data elements of variable size. The instructions also implement a synchronization mechanism that allows asynchronous execution of data move and compute operations across different components of the neural network accelerator as well as between multiple instances of the neural network accelerator.

Apparatuses, systems, and techniques to perform an operation to indicate one or more non-zero values within one or more matrices of data; to perform an API to compress one or more matrices of data; to perform a matrix multiply accumulate (MMA) operation on two or more matrices of data, wherein at least one of the two or more matrices contain compressed data; and/or to perform an API to decompress one or more matrices of data. In at least one embodiment, one or more circuits are configured to receive and compile one or more instructions to perform computational operations for a sparse matrix multiplication.

A trace subsystem of an emulation system may generate differential frame data based upon successive frames. If one compression mode, the trace subsystem may set a flag bit and store differential frame data if there is at least one non-zero bit in the differential frame data. If the differential frame data includes only zero bits, the trace subsystem may set the flag bit without storing the frame data. In another compression mode, the computer may further compress the differential data if the frame data includes one (one-hot) or two (two-hot) non-zero bits. The controller may set flag bits to indicate one of all-zeroes, one-hot, two-hot, and random data conditions (more than two non-zero bits). For one-hot or two-hot conditions, the controller may store bits indicating the positions of the non-zero bits. For random data conditions, the controller may store the entire differential frame.

A system and method for creating firmware patch files is disclosed. The method utilizes the Executable Linkable Format file that is created when the firmware image is created. By analyzing the ELF file, the patch creation software is able to identify functions and other data in the new firmware image. The patch creation software then compares these functions to corresponding functions in the old firmware image. The method then creates an edit sequence that may be used to transform the old firmware image into the new firmware image. The edit sequence is then converted into a series of opcodes, where each opcode is followed by at least one parameter. A patch program, disposed on a network device, is able to apply the patch file to update its firmware. This method creates a smaller patch file than other popular tools.

Disclosed are a computer system, a software path prediction computer, non-transitory computer-readable medium, and method for determining a predicted software path that utilize segmentation machine learning in combination with ensemble machine learning to keep a most accurate model running on a server program that receives requests from and sends predicted software path(s) to a software client.

A computing system and a compressing method for neural network parameters are provided. In the method, multiple neural network parameters are obtained. The neural network parameters are used for a neural network algorithm. Every at least two neural network parameters are grouped into an encoding combination. The number of neural network parameters in each encoding combination is the same. The encoding combinations are compressed with the same compression target bit number. Each encoding combination is compressed independently. The compression target bit number is not larger than a bit number of each encoding combination. Thereby, the storage space can be saved and excessive power consumption for accessing the parameters can be prevented.