2D/3D FPGA array for brain process and numerical computation
Jiang Li, Takahashi Ken’ichi, Hakaru Tamukoh, Masatoshi Sekine · 2012
We propose an FPGA Array structure that is designed to the scalable structure for executing brain processes and numerical computation for 3-D problem, and the circuits in FPGA Array are controlled from the application program running in host PC directly. Many circuits for brain processes are designed to evaluate the 3-D FPGA, also we implement arithmetic circuits that calculate 2D Poisson equation by the finite difference method to shows numerical computation performance. We also show the results of circuits for the brain processes such as voice recognition, voice synthesis, and image recognition with Self Organizing Map. We estimate the gate count of the circuit for practical brain processes.