摘要:
Novel and useful system and methods of functional safety mechanisms for use in an artificial neural network (ANN) processor. The mechanisms can be deployed individually or in combination to provide a desired level of safety in neural networks. Multiple strategies are applied involving redundancy by design, redundancy through spatial mapping as well as self-tuning procedures that modify static (weights) and monitor dynamic (activations) behavior. The NN processor incorporates functional safety concepts which reduce its risk of failure that occurs during operation from going unnoticed. The mechanisms function to detect and promptly flag and report the occurrence of an error with some mechanisms capable of correction as well. The safety mechanisms cover data stream fault detection, software defined redundant allocation, cluster interlayer safety, cluster intralayer safety, layer control unit (LCU) instruction addressing, weights storage safety, and neural network intermediate results storage safety.
摘要:
A method for determining two bits errors in transmission of 128 bits and the device for realization of this method is provided. By the method and device, the two error bits transferred bits can be determined and corrected by using least bits in operation. Therefore, the amount of data in transmission is increased with a least quantity and thus the transmission quality is not affected.
摘要:
A writing circuit for writing write data into a memory comprises an evaluator configured for providing an error handling code on the basis of the write data. A modifier reversibly modifies extended write data comprising both the write data and the error handling code in dependence on address information related to a writing address in order to provide modified extended write data. A writer writes the modified extended write data in a position of the memory defined by a writing address. A reading circuit for reading extended read data from a memory comprises a reader configured for reading the extended read data from a position of the memory defined by a reading address. A de-modifier modifies the extended read data in dependence on address information related to a reading address in order to provide extracted read data and an extracted error handling code. An error-detector detects based on the extracted error handling code whether the extracted read data comprises an error.
摘要:
Certain exemplary aspects of the present disclosure are directed towards methods and apparatuses in which logic circuitry generates an error detection code based on user data received from a host, and further generates a first set of check bits, to be written to the non-volatile memory circuit in conjunction with the user data, by combining the error detection code with a hashed data address of the user data. In some embodiments, the check bits associated with the user data providing verification that the user data was written in the appropriate physical block address of the non-volatile memory circuit.
摘要:
A method for demodulating a received signal relating to a sequence of transmitted symbols that have been modulated by continuous phase modulation includes normalizing samples of a sequence of samples generated from the received signal, to obtain a normalized sequence of samples, wherein an amplitude of each sample of the normalized sequence of samples has an absolute value equal to unity; estimating, on the basis of the normalized sequence of samples, a time offset and a frequency offset of the received signal, and using the estimated time offset and the estimated frequency offset for compensating the normalized sequence of samples for the time and frequency offsets, to obtain a compensated sequence of samples; and determining a sequence of symbols corresponding to the transmitted sequence of symbols on the basis of the compensated sequence of samples. Also disclosed is a receiver for demodulating a received signal relating to a sequence of transmitted symbols that have been modulated by continuous phase modulation.
摘要:
A data transmission system comprising an Automotive Sensor Network System (ASNS) connected to a plurality of source locations via a common bus, wherein the ASNS is configured to ascertain the source from which the data-frames and first package checksum are received and based on the ascertainment of the source, appropriate decoding methods are used to calculate the ASNS location data-frame checksums and the ASNS location package checksums. A higher order redundancy check is done over a series of data-frames to detect errors in the reception caused by temporary high interference that may exist in the transmission path.
摘要:
In one embodiment, the present invention includes a method for receiving incoming data in a processor and performing a checksum operation on the incoming data in the processor pursuant to a user-level instruction for the checksum operation. For example, a cyclic redundancy checksum may be computed in the processor itself responsive to the user-level instruction. Other embodiments are described and claimed.
摘要:
Methods, apparatus and computer program products implement embodiments of the present invention that include replacing, in one or more initial source code files, each reference to a first function configured to convey system messages with a respective reference to a second function configured to convey the system messages, thereby producing respective corresponding preprocessed source code files for the one or more initial source code files. The respective corresponding preprocessed source code files are then compiled, thereby creating an executable file. While executing the executable file, a call to the second function is received, wherein the call includes a text string. A name of one of the respective corresponding preprocessed source code files storing the call to the second function is identified, and based on the identified name and the text string, a computed destination is determined for the text string. Finally, the text string is conveyed to the computed destination.
摘要:
Various aspects directed towards facilitating an error detection optimization over a shared bus are disclosed. A master device is coupled to a slave device, and an encoded communication of a word is facilitated between the master device and the slave device via a control data bus. The encoded communication is encoded according to a protocol that allocates a plurality of least significant bits of the encoded communication to facilitate maximizing an error detection constant. The protocol allocates the plurality of least significant bits to include at least one additional error detection bit or at least a first most significant bit of a data portion of the word.
摘要:
Format preserving encryption (FPE) cryptographic engines are provided for performing encryption and decryption on strings. A plaintext string may be converted to ciphertext by repeated application of a format preserving encryption cryptographic algorithm. Following each application of the format preserving cryptographic algorithm, the resulting version of the string may be analyzed to determine whether desired string constraints have been satisfied. If the string constraints have not been satisfied, further applications of the format preserving cryptographic algorithm may be performed. If the string constraints have been satisfied, the current version of the string may be used as an output for the cryptographic engine.