High-Level Development Framework

NS-TCL™ is a high-level development framework based on OpenCL™ for writing executable software across heterogeneous platforms including field programmable gate arrays FPGAs, central processing units CPUs, graphics processing units GPUs, and digital signal processors DSPs.

Trending research and developments on FPGA based model accelerations for Deep Learning have advanced methods for implementing model layers and features of Convolutional Neural Networks. The rapid developments of this research have been initiated by OpenCL™ (Open Computing Language). OpenCL™ enables high-level programming for several processing architectures.

FPGA based Deep Learning model accelerations provide large benefits for scalable applications which require low energy consumption rates. Though several frameworks have exploited FPGA accelerations of Deep Learning on single architectures, NS-TCL™ leverages the collaborated performance of distributed FPGA architectures on distributed systems such as “AI Smart Classrooms” which require correlative processing of data form sensors and other input devices to perform a series of tasks.

Our heterogeneous platform development framework explores distributed FPGA and SoC software implementations in Edge based Deep Learning systems for smart and resillient IoT Networks.

#NS-T for Acceleration