Hybrid Spatial and Temporal Computing Histogrammer in Soft Processor Core of a FPGA Device

Multi-channel data management is crucial in a world where big data processing is extensively used in research and business. Histogramming is a common technique employed to detect, analyze, and store enormous volumes of data in real-time, making it useful for industrial applications in fields such as biology, chemistry, medical imaging, and spectroscopy. Due for them programming simplicity and low-cost large amount of memory, general-purpose temporal computing processors are commonly used, but they lack the ability to perform parallel computation at high-performance. Field-Programmable Gate Array (FPGA) is a powerful parallel computing solution proposed by both the scientific and industrial worlds, but it is equipped with little memory for these applications. Thus, a hybrid spatial/temporal computing histogram generator has been proposed, which uses a low-area multi-channel histogramming engine in programmable logic which is expanded thanks to an external Double Data Rate Synchronous Dynamic Random Access Memory (DDR) driven by a MicroBlaze Soft Processor Core. The proposed system has been validated on a Xilinx 28-nm 7-Series Artix-7 XC7A100T FPGA hosted on a Nexys4 Evaluation Board. Thanks to this hybrid solution, up to 128 channels can handle in a low-end FPGA occupies 207 LUTs and 325 flip-flops per channel plus a total 630 kb of total BRAM shared between all channels; a power consumption of 10.1 mW per channel is measured.