@inproceedings{3b1d788d8f054f05bd5706698f5bb31a,
title = "FlexiDRAM: A Flexible in-DRAM Framework to Enable Parallel General-Purpose Computation",
abstract = "In this paper, we propose a Flexible processing-in-DRAM framework named FlexiDRAM that supports the efficient implementation of complex bulk bitwise operations. This framework is developed on top of a new reconfigurable in-DRAM accelerator that leverages the analog operation of DRAM sub-arrays and elevates it to implement XOR2-MAJ3 operations between operands stored in the same bit-line. FlexiDRAM first generates an efficient XOR-MAJ representation of the desired logic and then appropriately allocates DRAM rows to the operands to execute any in-DRAM computation. We develop ISA and software support required to compute in-DRAM operation. FlexiDRAM transforms current memory architecture to a massively parallel computational unit and can be leveraged to significantly reduce the latency and energy consumption of complex workloads. Our extensive circuit-to-architecture simulation results show that averaged across two well-known deep learning workloads, FlexiDRAM achieves ~15 energy-saving and 13 speedup over the GPU outperforming recent processing-in-DRAM platforms.",
author = "Ranyang Zhou and Arman Roohi and Durga Misra and Shaahin Angizi",
note = "Funding Information: ACKNOWLEDGMENTS This work is partially supported by a National Science Foundation grant (#ECCS-1710009). REFERENCES Funding Information: 5 CONCLUSION In this paper, we presented the FlexiDRAM framework to support the efficient implementation of complex bulk bitwise operations in DRAM. FlexiDRAM generates an XOR-MAJ representation of the desired logic and appropriately allocates DRAM rows to the operands to execute any in-DRAM computation. The framework is supported by ISA and the interface required to compute in-DRAM operation. Our results demonstrate that averaged across two deep learning workloads, FlexiDRAM achieves ∼15× energy-saving and 13× over the GPU, outperforming recent in-DRAM accelerators. Publisher Copyright: {\textcopyright} 2022 Copyright held by the owner/author(s).; 2022 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2022 ; Conference date: 01-08-2022 Through 02-08-2022",
year = "2022",
month = aug,
day = "2",
doi = "10.1145/3531437.3539721",
language = "English (US)",
series = "Proceedings of the International Symposium on Low Power Electronics and Design",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2022 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2022",
address = "United States",
}