Multiarchitecture Hardware Acceleration of Hyperdimensional Computing with oneAPI

Hyperdimensional computing (HDC) is a machine-learning method that seeks to mimic the high-dimensional nature of data processing in the cerebellum. This research utilizes oneAPI libraries with SYCL to create multiple accelerators for HDC learning tasks for CPUs, GPUs, and field-programmable gate arrays (FPGAs). This work benchmarked HDC training and inference tasks on the Intel Xeon Platinum 8256 CPU, Intel UHD 11th generation GPU, and Intel Stratix 10 FPGA. The GPU implementation had the fastest training times and highest throughput. The FPGA implementation had the lowest inference latency.

Download Presentation