SMART-WRITE: Adaptive Learning-based Write Energy Optimization for Phase Change Memory
Mahek Desai, Rowena Quinn, Marjan Asadinia
随着动态随机访问存储器(DRAM)和其他基于晶体管的内存接近其可扩展性极限,寻找替代存储方法变得越来越紧迫。 与许多现有内存技术相比,相变存储器(PCM)因其可扩展性、快速访问时间和零泄漏功率而成为有前途的候选者。 然而,PCM存在重大缺陷,目前阻碍了其作为替代品的可行性。 PCM细胞的寿命有限,因为写入操作会降低物理材料,这些操作会消耗相当多的能量。 使PCM成为数据存储的实用选项 - 涉及频繁写入操作 - 必须增强其细胞耐力,并且必须降低写入能量。 在本文中,我们提出了SMART-WRITE,这是一种集成神经网络(NN)和强化学习(RL)的方法,以动态优化写入能量并提高性能。 NN模型监控实时操作条件和设备特性,以确定最佳的写入参数,而RL模型动态调整这些参数,以进一步优化PCM的能耗。 通过根据实时系统条件持续调整 PCM 写入参数,SMART-WRITE 可将写入能耗降低高达 63%,与基线和前型号相比,性能提升高达 51%。
As dynamic random access memory (DRAM) and other current transistor-based memories approach their scalability limits, the search for alternative storage methods becomes increasingly urgent. Phase-change memory (PCM) emerges as a promising candidate due to its scalability, fast access time, and zero leakage power compared to many existing memory technologies. However, PCM has significant drawbacks that currently hinder its viability as a replacement. PCM cells suffer from a limited lifespan becau...