An Energy-Efficient In-Memory Accelerator for Graph Construction and Updating
Mingkai Chen,Cheng Liu,Shengwen Liang,Lei He,Ying Wang,Lei Zhang,Huawei Li,Xiaowei Li
DOI: https://doi.org/10.1109/tcad.2024.3355038
IF: 2.9
2024-01-01
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Abstract:Graph is widely utilized as a key data structure in many applications such as social network and recommendation systems. However, many real-world graphs are constructed with large-scale correlated data, which typically involves massive random memory accesses and distance calculation resulting in considerable processing time and energy consumption on CPUs and GPUs. In this work, we present GCiM, a specialized processing-in-memory architecture for efficient graph construction and updating. By directly deploying the computing units on the logic layer of the 3D stacked memory, GCiM benefits from memory-level parallelism and further improves the memory access efficiency with both optimized processing ordering and data layout. In addition, we notice that the computing engines for graph updating suffer from dramatic utilization imbalance and integrate a power gating module to cut down the power supply of the idle computing engines at runtime and further enhance the energy efficiency. According to our experiments, GCiM shows 634.64X and 56.27X speedup while consuming 1194.14X and 505.07X less energy compared to CPU and GPU respectively.
engineering, electrical & electronic,computer science, interdisciplinary applications, hardware & architecture