With the increase of available storage bandwidth, CPUs can not keep up with the compute throughput needed to process this amount of incoming data. GPUs and FPGAs are generally better suited for such tasks. To assist FPGAs in their functions, some boards are equipped with one or more high bandwidth memory (HBM) stacks, with a bandwidth of 230 GB/s each. This thesis presents a hardware design for the Alveo U280 FPGA board with HBM. Each HBM stack provides multiple interfaces to the full range of memory within HBM. Utilizing these multiple interfaces, a hardware decompressor for the Snappy compression algorithm is placed in parallel to achieve a higher end-to-end throughput. Additionally a design is created to perform benchmarks on HBM where varying sizes of data are transported between HBM and logic within the FPGA. The hardware decompressor and component that interfaces to memory within HBM were found to be incompatible and required additional logic to become able to transport data between them. To ease the parallelization of the decompressor a custom Snappy framing format is implemented. Using this format a softcore processor on the FPGA is able to buffer the locations of compressed data within HBM and divide these over available decompressors. The design is successfully synthesized into a kernel that can be loaded by the FPGA. From the moment compressed data is sitting in HBM until it is decompressed, a single decompressor reaches a maximum end-to-end throughput of 4.0 GB/s. When eight or more decompressors are activated, they reach a throughput between 20.0 to 26.2 GB/s. The hardware decompressor designs uses less than 10% of the resources of the U280 with little power usage. Compared to a software implementation, using multithreading, the hardware solution is 1.5-2.5x faster on a set of files that is used for benchmarks on decompression speed with varying compression ratios.

Pressure ulcers develop when skin is subjected to a mechanical load over time, resulting in tissue damage. This occurs when a patient is in the same posture or position for a long period of time. To prevent this Momo Medical is developing a sensor plate that uses sensor data attained from the patient laying on his bed to guide the nurses, whom are responsible for re-positioning the patients. The sensor data is processed by an algorithm to detect the patient and determine the position of the patient. The sensor plate is placed between the mattress and the frame and has sensors to detect the dynamic and static behaviour of the patient.The goal of this project as a group is to facilitate the performance of the Momo medical system. This is done by studying the currently used system to find ways to improve it or find better alternatives for specific parts of the system and by giving recommendations based on the outcome of this project. The group was divided into three subgroups. Two subgroups focus on the two different types of sensors used to detect the dynamic and static behaviour of the patient. The last group facilitates the Momo system patient detection algorithm by detecting a patient’s heart rate from pressure sensor data using a microcontroller-implemented algorithm. The data received is also displayed real-time on a user-interface to aid the testing and validation of the system. The designed algorithm uses wavelet transformed peak detection to analyze the data for a ballistocardiography heart signal. The wavelet transform is used to extract the ballistocardiography heart signal from the sensor data and the peak detection is then used to find the specific heartbeats present. The algorithm was simulated and tested in MATLAB and this resulted in a less than 10% error compared to an electrocardiography based reference signal, that was measured using the Polar T34. The different functions needed for the algorithm were implemented on the microcontroller and are functional. Also the interface used to display the incoming sensor data is functional. The actual implementation and testing of the algorithm on the microcontroller has not been finished yet. However based on the design of the algorithm in MATLAB and the function implementations that already work on the microcontroller, the detection of a patient’s heart rate from pressure sensor data using a microcontroller-implemented algorithm should be entirely feasible.