PHASECHANGE TECHNOLOGY AND THE （相变技术和）.pdfVIP

PHASE-CHANGE TECHNOLOGY AND THE FUTURE OF MAIN MEMORY PHASE-CHANGE MEMORY MAY ENABLE CONTINUED SCALING OF MAIN MEMORIES, BUT PCM HAS HIGHER ACCESS LATENCIES, INCURS HIGHER POWER COSTS, AND WEARS OUT MORE QUICKLY THAN DRAM. THIS ARTICLE DISCUSSES HOW TO MITIGATE THESE LIMITATIONS THROUGH BUFFER SIZING, ROW CACHING, WRITE REDUCTION, AND WEAR Benjamin C. Lee LEVELING, TO MAKE PCM A VIABLE DRAM ALTERNATIVE FOR SCALABLE MAIN MEMORIES. Stanford University Over the past few decades, mem- Programming mechanisms such as current Ping Zhou ory technology scaling has provided many injection scale with cell size. Phase-change Jun Yang benefits, including increased density and ca- memory (PCM), spin-torque transfer pacity and reduced cost. Scaling has provided (STT) magnetoresistive RAM (MRAM), Youtao Zhang these benefits for conventional technologies, and ferroelectric RAM (FRAM) are examples such as DRAM and flash memory, but of resistive memories. Of these, PCM is clos- Bo Zhao now scaling is in jeopardy. For continued est to realization and imminent deployment scaling, systems might need to transition as a NOR flash competitor. In fact, various University of Pittsburgh from conventional charge memory to emerg- researchers and manufacturers have proto- ing resistive memory. Charge memories re- typed PCM arrays in the past decade.2 quire discrete amounts of charge to induce PCM provides a nonvolatile storage Engin Ipek a voltage, which is detected during reads. mechanism that is amenable to process scal- In the nonvolatile space, flash memories ing. During writes, an access transistor injects University of Rochester must precisely control the disc