← 返回 JSSC 论文列表JSSC 2024第6期MemorySRAMCIM
Cramming More Weight Data Onto Compute-in- Memory Macros for High Task-Level Ene
提出一种基于1T-MLC ROM的存内计算宏单元,以提高存储密度和能效。
未明确提及具体性能指标
存内计算静态随机存取存储器多级单元只读存储器能量效率
▸创新点1:多源驱动1T-MLC ROM(电路创新) - 通过多源驱动技术实现单晶体管多级单元ROM,显著提升存储密度和能效比,解决了传统SRAM基CiM单元密度低的问题,支持多比特MAC操作。
▸创新点2:电荷域电容共享技术(电路创新) - 采用电荷域电容共享技术进一步压缩单元面积,实现超高密度CiM存储,相比传统方案存储密度提升显著,具体指标未在摘要中提及但通过技术描述可推断。
▸创新点3:基于ROM的迁移学习架构(系统创新) - 设计ROM基迁移学习架构,灵活支持不同任务且精度损失极小,通过硬件-算法协同优化提升任务级能效,适用于数据密集型场景。
▸创新点4:1T-MLC ROM基CiM宏单元(方法创新) - 首次提出将1T-MLC ROM用于CiM宏单元,结合多级存储和计算一体化设计,为高能效边缘计算提供新范式。
Abstract
Owing to the mature process and low access
energy, static random-access memory (SRAM) has become a
promising candidate for compute-in-memory (CiM) acceleration
of multiply-accumulate (MAC) operations. However, SRAM-
based CiM cells have rather low density and thus very limited
total on-chip memory capacity. This fact, unfortunately, results
in undesired weight data reload operations from the off-chip
dynamic random-access memory (DRAM) in data-intensive sce-
narios and may even tarnish the energ