Abstract: A request processing method, includes: acquiring, from a local first input first output (FIFO) queue, an address of a current request to be processed, and based on the address of the current request to be processed, acquiring the current request to be processed from a preset request buffer area; parsing the current request to be processed, and based on a parsing result, acquiring corresponding current source data to be processed and a serial address item for recording an address of a next request to be processed; processing the current source data to be processed to obtain currently processed target data, and writing the currently processed target data into a preset data buffer area; and notifying a host to read all processed target data from the preset data buffer area when the serial address item is empty.

Abstract: Certain examples described herein relate to a co-processor that allows a processing unit to efficiently perform a cryptographic operation. The co-processor has an arithmetic unit that is configured to perform discrete binary arithmetic using bit sequences loaded from a memory. The co-processor may be configured for fast, low-power computation of certain functions that comprise low-level building blocks for the cryptographic operation. These functions may include Boolean logic and integer arithmetic. The co-processor has a set of control registers that are writable by the processing unit to control the co-processor. Addresses for one or more sources and destinations may be computed by the co-processor to allow for flexible operation. The co-processor may allow many advanced cryptographic operations to be rapidly computed, including those that are “post-quantum” secure.

Abstract: Embodiments detailed herein relate to matrix operations. In particular, support for a matrix transpose instruction is detailed. In some embodiments, decode circuitry to decode an instruction having fields for an opcode, a source matrix operand identifier, and a destination matrix operand identifier; and execution circuitry to execute the decoded instruction to transpose each row of elements of the identified source matrix operand into a corresponding column of the identified destination matrix operand are detailed.

Abstract: An exemplary register circuit includes a plurality of slots to store respective addresses and data pairs. During a write operation, each slot of a plurality of slots preceding a particular slot of the plurality of slots indicated as empty is shifted by one slot to fill the particular slot such that a first end slot of the plurality of slots is made available to receive a new write address and data pair. Each slot of the plurality of slots subsequent to the particular slot retains existing address and data pairs.

Abstract: A processor of an aspect includes a decode unit to decode a matrix multiplication instruction. The matrix multiplication instruction is to indicate a first memory location of a first source matrix, is to indicate a second memory location of a second source matrix, and is to indicate a third memory location where a result matrix is to be stored. The processor also includes an execution unit coupled with the decode unit. The execution unit, in response to the matrix multiplication instruction, is to multiply a portion of the first and second source matrices prior to an interruption, and store a completion progress indicator in response to the interruption. The completion progress indicator to indicate an amount of progress in multiplying the first and second source matrices, and storing corresponding result data to the third memory location, that is to have been completed prior to the interruption.

Abstract: A non-transitory computer-readable recording medium stores a program for causing a computer to execute a process, the process includes determining, for an n-dimensional array (n?3) included in an instruction code in an innermost loop of a multiple loop included in a source code, whether array sizes of a first argument and a second argument match numbers of rotations of a first index and a second index in the multiple loop, respectively, when the array sizes match the numbers of rotations and when each initial value and each increment value of the first and second indexes is 1, replacing the first argument and the second argument of the n-dimensional array included in the instruction code with a third argument and changing the n-dimensional array to an (n?1)-dimensional array, and integrating a loop that uses the first index and a loop that uses the second index.

Abstract: A memory device includes a command decoder that implements security logic to detect a command sequence to read a security region of a memory array with continuous encrypted data and to output/input specific contexts for the data. Output/input of specific contexts can be during a dummy cycle to achieve greater performance. A host interfacing can, for example, execute a single command to both get the encrypted data and specific contexts that were used to encrypt the data. Our technology can implement transferring data on the system bus in ciphertext and encrypted by a different Nonce or a different session key than used in a previous transfer operation. In this way, data will be represented with different ciphertext on the bus at different sessions; thereby defending against a replay attack.

Abstract: An arithmetic device includes an AF circuit including a first table storage circuit. The AF circuit stores a table input signal into one variable latch selected based on an input selection signal among variable latches included in the first table storage circuit in a look-up table form when a table set signal is activated. The AF circuit extracts a result value of a first activation function realized by a look-up table based on an input distribution signal to output the extracted result value as a fist table output signal for generating an output distribution signal.

Abstract: An integrated circuit (“IC”) module includes a substrate, multiple field-programmable gate array (“FPGA”) dies, and pads capable of being selectively configured to perform one or more user defined logic functions. The substrate is configured to house multiple FPGA dies side-by-side in an array formation facilitating transmission of signals between the FPGA dies or chips. The FPGA dies are placed on the substrate functioning as a single FPGA device. The periphery dies of the FPGA dies are configured for external connectivity and the interior dies which are interconnected to perform user defined logic functions. The pads, in one aspect, coupling to the FPGA dies, are configured to provide connections between at least some of the FPGA dies.

Abstract: Systems, apparatuses and methods suitable for optimizing synchronization mechanisms for multi-core processors are provided. The synchronizing mechanisms may be optimized by receiving a command stream which comprises a plurality of commands including one or more wait commands, wherein each wait command has an associated state and one or more associated conditions; sequentially processing each command in the command stream until a wait command is reached; checking the state associated with the wait command to be processed, wherein if said state is a blocking state, further processing of commands in the command stream is paused until each of said wait command's associated conditions are met, and wherein if said state is a non-blocking state, the next command in the command stream is retrieved and processed.

Abstract: A device includes a matrix transpose component, a matrix processing component, a data alignment component, and a data reduction component. The matrix transpose component is configured to transpose an input matrix of elements to output an output matrix of the elements that have been transposed. The matrix processing component is configured to multiply a first multiplication input matrix with a second multiplication input matrix, wherein the output matrix of the matrix transpose component is utilized as the first multiplication input matrix and a mask vector is utilized as the second multiplication input matrix. The data alignment component is configured to modify at least a portion of elements of a result of the matrix processing component. The data reduction component is configured to sum at least the elements of the modified result of the matrix processing component to determine a sum of the group of values.

Abstract: Methods, systems, articles of manufacture, and apparatus are disclosed to decode zero-value-compression data vectors. An example apparatus includes: a buffer monitor to monitor a buffer for a header including a value indicative of compressed data; a data controller to, when the buffer includes compressed data, determine a first value of a sparse select signal based on (1) a select signal and (2) a first position in a sparsity bitmap, the first value of the sparse select signal corresponding to a processing element that is to process a portion of the compressed data; and a write controller to, when the buffer includes compressed data, determine a second value of a write enable signal based on (1) the select signal and (2) a second position in the sparsity bitmap, the second value of the write enable signal corresponding to the processing element that is to process the portion of the compressed data.

Abstract: A circuit interface includes one or more processors that generate opcodes, a plurality of interface control circuits, each including a respective processing element responsive to the opcodes generated by one or more processors. Each interface control circuit corresponds to a respective link of a plurality of links of a device-to-device interface (DDI), and each link of the plurality of links of the DDI is for transmitting or receiving signals from one or more sources or one or more destinations external to the circuit.

Abstract: This application discloses a pooling processing method, applied to a pooling processing system of a convolutional neural network. The pooling processing system includes a first storage device, a data region, a pooling computation kernel, and a pooling controller.

Abstract: Convolutions of an input sample with multiple kernels is decomposed into matrix multiplications of a V×C matrix of input values times a C×K matrix of kernel values, producing a V×K product. For the second matrix, C is a channel dimension (i.e., each row of the second matrix is a different channel of the input sample and kernel) and K is the kernel dimension (i.e., each column of the second matrix is a different kernel), but all the values correspond to the same pixel position in the kernel. In the matrix product, V is the output dimension and K is the kernel dimension. Thus, each value in the output matrix is a partial product for a certain output pixel and kernel, and the matrix multiplication parallelizes the convolutions by calculating partial products for multiple output pixels and multiple kernels.

Abstract: Methods and apparatus for ultra-secure accelerators. New ISA enqueue (ENQ) instructions with a wrapping key (WK) are provided to facilitate secure access to on-chip and off-chip accelerators in computer platforms and systems. The ISA ENQ with WK instructions include a dest operand having an address of an accelerator portal and a scr operand having the address of a request descriptor in system memory defining a job to be performed by an accelerator and including a wrapped key. Execution of the instruction writes a record including the src and a WK to the portal, and the record is enqueued in an accelerator queue if a slot is available. The accelerator reads the enqueued request descriptor and uses the WK to unwrap the wrapped key, which is then used to decrypt encrypted data read from one or more buffers in memory. The accelerator then performs one or more functions on the decrypted data as defined by the job and writes the output of the processing back to memory with optional encryption.

Abstract: An integrated circuit (“IC”) module includes a substrate, multiple field-programmable gate array (“FPGA”) dies, and pads capable of being selectively configured to perform one or more user defined logic functions. The substrate is configured to house multiple FPGA dies side-by-side in an array formation facilitating transmission of signals between the FPGA dies or chips. The FPGA dies are placed on the substrate functioning as a single FPGA device. The periphery dies of the FPGA dies are configured for external connectivity and the interior dies which are interconnected to perform user defined logic functions. The pads, in one aspect, coupling to the FPGA dies, are configured to provide connections between at least some of the FPGA dies.

Abstract: In some examples, with respect to hash offset based key version embedding, data that is to be encrypted may be ascertained, and a key, including a key version, that is to be used to encrypt the ascertained data may be ascertained. Encrypted data may be generated by encrypting the ascertained data based on the ascertained key, and hashed encrypted data may be generated by performing a hash operation on the encrypted data. Further, offset hashed encrypted data may be generated by embedding the key version into the hashed encrypted data, and the offset hashed encrypted data including the embedded key version may be stored.

Abstract: A method and system for monitoring computer network intrusions, the system comprising at least one security device including a processor and memory. The at least one security device is communicatively coupled to a private network and configured to generate heartbeat pulses comprising operational snapshots of the at least one security device. The system further comprises one or more host systems configured to communicate with the at least one security device from an external network, transmit configuration parameters to the at least one security device, the configuration parameters including instructions for the at least one security device to operate as a given type of network asset, monitor the heartbeat pulse of the at least one security device, determine a change in integrity in the at least one security device based on the monitoring, and send one or more notification messages to a network administrator based on the determination.

Abstract: The invention disclosed herein is an in-circuit security system for electronic devices. The in-circuit security system incorporates identity credential verification, secure data and instruction storage, and secure data transmission capabilities. It comprises a single semiconductor chip, and is secured using industry-established mechanisms for preventing information tampering or eavesdropping, such as the addition of oxygen reactive layers. This invention also incorporates means for establishing security settings, profiles, and responses for the in-circuit security system and enrolled individuals. The in-circuit security system can be used in a variety of electronic devices, including handheld computers, secure facility keys, vehicle operation/ignition systems, and digital rights management.

Abstract: A host writes graphics commands and data to programmable registers through a command FIFO that is read by a graphics controller or BitBlt engine. Rather than write an address and a data value for each register programmed, the host writes one address, one index, and several data values. The address points to an index register. The index is a mapping index word with several multi-bit mapping fields. Each multi-bit mapping field in the index identifies a register to be programmed with one of the data values. Since N bits are used for each mapping field, the mapping field can select one register in a bank of 2N?1 registers. The registers in the bank can be programmed in any order, and registers can be skipped. Since only one index is stored in the command FIFO for programming several registers, less memory space and fewer bus cycles are required.