High Performance Computing (HPC) is a hardware parallel computation acceleration technology that delivers performance well exceeding the computation power of the regular computer software. Graphics Processing Unit (GPU) and Field Programming Gate Array (FPGA) are the two common technologies used in HPC. GPU is original designed to process image data and includes large amount of processing cores for parallel processing. FPGA is a programmable logic circuit chip. Many of FPGA chips also have hardware DSP units for hardware computation. Although the computation power of FPGA is less than that of GPU, it has the advantage of integrating other customized digital circuits in one chip that the GPU is not capable of. Thus FPGA is the better technology for HPC in measurement and control devices and make device on the chip possible.
Our lab has extended the expertise in FPGA applications to HPC application in medical devices and is equipped with the National Instruments LabVIEW FPGA and FlexRIO hardware dedicated to HPC research projects.
Current project applied the HPC using FPGA in real-time fiber optic monitoring for Spinal Cord Ischemia. The huge amount of data generated from the optical device requires high computation power for time critical analysis. Such power is not available in conventional software approach. Our solution is to use FPGA to realize the computation of diffusive correlation spectroscopy model. We use LabVIEW FPGA development platform from National Instruments with FPGA hardware modules. We adopted the PXI chassis that can seamlessly integrate the FPGA modules with PC in a portable form factor. The portable real time spinal cord ischemia monitor is based on the technology and will have many medical applications.