Designed from the ground up to address significant machine learning (ML) inference deployment challenges facing system on chip (SoC) developers, Quadric's Chimera™ General Purpose Neural Processing Unit (GPNPU) family has a simple yet powerful architecture with demonstrated improved matrix-computation performance over the traditional approach. Its crucial differentiation is its ability to execute diverse workloads with great flexibility all in a single processor scaling from 1 TOP to 864 TOPs.