Abstract: An edge device for image processing includes a series of linked components which can be independently optimized. A specialized change detector which optimizes the events collected at the expense of false positives is accompanied by a trainable module, which uses training feedback to reduce the false positives over time. A “look ahead module” peeks ahead in time and determines whether an inference pipeline needs to run. This allocates a definite amount of time for the validation and training module. The training module is operated in terms of a quantum of time. Processing time during phases of no scene activity is reserved to carry out training. A lightweight detector and the classifier are trainable modules. A site optimizer is made up of rules and sub-modules using spatio-temporal heuristics to handle specific false positives while optimally combining the change detector and inference module results.

Abstract: A novel coprocessor interface providing memory access without traversing the main processor, and methods of operating the same. A system includes a bus, a processor circuit, a memory circuit, a multi-channel memory controller, and at least one coprocessor. The processor circuit is coupled to the bus, the multi-channel memory controller is coupled between the bus and the memory circuit, and the coprocessors are coupled to both the processor circuit and the multi-channel memory controller. This circuit arrangement provides dedicated high speed channels for data access between the coprocessors and the memory circuit, without traversing the processor circuit or the bus. Thus, non-standard (e.g., non-sequential) data transfer protocols can be supported. In some embodiments, the system is implemented in a programmable logic device (PLD). The processor circuit can be, for example, a microprocessor included as hard-coded logic in the PLD, or can be implemented using programmable logic elements of the PLD.

Abstract: Various approaches for profiling a target system are described. In one approach, a uni-directional, point-to-point bus has a single input port and a single output port. A target processor has a trace port coupled to the input port of the bus and is configured to execute a plurality of instructions one or more times. The target processor provides state data at the trace port and to the input port of the bus. A profile circuit arrangement is coupled to the output port of the first bus, and a memory is coupled to the profile circuit arrangement. The profile circuit arrangement is configured to read data from the output port of the first bus and write the data to the memory.

Abstract: A method of recreating instructions and data traces in a processor can include the step of fetching an instruction from an executable program in an order corresponding to sequential program counter (PC) values, obtaining a destination register from the fetched instruction and updating the destination register in a data structure with a value from a collected destination register corresponding to the PC value. The steps above can be repeated until all desired PC values and destination values are obtained.

Abstract: A method of implementing a comparator in a device having programmable logic is described. The method comprises implementing a first comparison function in a first lookup table; implementing a second comparison function in a second lookup table; and using an output associated with the first comparison function to select an output of the comparator. A device having programmable logic comprising a comparator is also described.