Carnegie Mellon University

Franz Franchetti

Lawrence Thomas Pileggi

Qiuling Zhu

Larry Pileggi

Fazle Sadi

Disclosed herein is a novel multi-way merge network, referred to herein as a Hybrid Comparison Look Ahead Merge (HCLAM), which incurs significantly less resource consumption as scaled to handle larger problems. In addition, a parallelization scheme is disclosed, referred to herein as Parallelization by Radix Pre-sorter (PRaP), which enables an increase in streaming throughput of the merge network. Furthermore, high performance reduction scheme is disclosed to achieve full throughput.

High performance merge sort with scalable parallelization and full-throughput reduction

This disclosure relates to a three-dimensional (3D) integrated circuit (3DIC) memory chip including computational logic-in-memory (LiM) for performing accelerated data processing. Related memory systems and methods are also disclosed. In one embodiment, the 3DIC memory chip includes at least one memory layer that provides a primary memory configured to store data. The 3DIC memory chip also includes a computational LiM layer. The computational LiM layer is a type of memory layer having application-specific computational logic integrated into local memory while externally appearing as regular memory. The computational LiM layer and the primary memory are interconnected through through-silica vias (TSVs). In this manner, the computational LiM layer may load data from the primary memory with the 3DIC memory chip without having to access an external bus coupling the 3DIC memory chip to a central processing unit (CPU) or other processors to computationally process the data and generate a computational result.

Growing community of inventors

Pittsburgh, PA, United States of America

Franz Franchetti