Skip to main content
. 2016 Jan 14;9:516. doi: 10.3389/fnins.2015.00516

Figure 3.

Figure 3

Compilation pipeline and runtime file loading of NeuroFlow. The pipeline receives neural model code written in Python which is then compiled into hardware configuration, host side configuration and memory data file. The resulting data files and configurations are loaded into the host processor (CPU) and FPGA of the system during runtime.