摘要:
In the paper GPU Computing by J. D. Owens, M. Houston et al., 2008 the authors stated that the graphics processing unit (GPU) has become an integral part of today's mainstream computing systems[1]. Seven years later the GPU is going to become an integral part of embedded computing. The ongoing automation in industrial fields and automated driving made image processing a key workload for autonomous systems. So combining a CPU and GPU in one system on chip (SOC) is a useful step to reduce consumption and cost of these systems. The NVIDIA Tegra K1 is the first SOC with a CUDA capable general purpose GPU (GPGPU) and an energy efficient ARM CPU to fulfill these parallel workloads and high demands on efficient energy consumption and high performance.This paper gives an insight into the SOC and examines whether it can be used as a lower power high performance System. Further- more, the chip is compared with an existing high performance system and the results are interpreted.
展开