增强RRAM可靠性的热通量压缩算法

为提高阻变存储器(RRAM)的可靠性,研究了RRAM中的热串扰问题,提出了一种热通量压缩(TFC)算法,通过在RRAM读写电路之前加入TFC算法,降低在RRAM中产生的焦耳热,从而减弱RRAM中的热串扰问题,提高RRAM的可靠性。TFC算法通过分析计算写入数据流产生的真实焦耳热热量进而判断是否对写入数据流进行翻转编码,即TFC算法会在原始写入数据流和翻转编码数据流二者中选择热通量较小者通过算法层,从而达到减小RRAM存储器中焦耳热热量的目的。理论分析和仿真结果表明:以字节为单位数据块条件下TFC算法平均可降低30%以上的写入焦耳热,阻变单元的保持时间估值平均增加了35%以上。 

Thermal flux compression algorithm to enhance RRAM reliability

A thermal flux compression (TFC) algorithm is proposed through the research on the thermal crosstalk in resistive random access memory (RRAM) to strengthen the RRAM reliability. By adding TFC algorithm before RRAM read-write circuit and reducing Joule heat produced in RRAM, thermal crosstalk in RRAM could be weakened and RRAM reliability could be enhanced. With the analysis and computation of real Joule heat produced in write data flow, TFC algorithm could judge whether to flip data, in order to reduce the Joule heat in RRAM, that is, TFC algorithm would choose the one with lower thermal flux between original write data flow and flipped encode data flow through the algorithm layer. Theoretical analysis and simulation results show that TFC algorithm could reduce more than 30% write Joule heat on average and valuation of resistive unit retention could increase by over 35% on average, with byte as the unite data block.