Abstract

Main memory is acritical component of all computer systems. Memory system must scale in size, cost and performance to maintain overall computer performance growth, but there are a lot of challenges that memory systems have encountered and throughout recent years some (potentially temporary) limitations for memory systems were reached out. This paper will discuss recent challenges in main memory systems and solutions that can be proposed to address those challenges. Problems to be discussed in the paper: memory capacity, energy consumption, periodic refresh, scaling, memory interference, “Rowhammer”.

References

• LimK., Jichuan C., Mudge T., Ranganathan P., Reinhardt S., Wenisch T. 2009. Disaggregated Memory for Expansion and Sharing in Blade Servers. ISCA '09 2009 Computer Science.
• Mutlu, O., Meza, J., Subramanian, L. 2015. The Main Memory System: Challenges and Opportunities. Communications of the Korean Institute of Information Scientists and Engineers, 33, 16-41.
• Lefurgy C., Rajamani K., Rawson F., Felter W., Kistler M. Tom K. 2003. "Energy Management for Commercial Servers", IEEE Computer, pp. 39-48, December, 2003.
• Li Y., Lefurgy C., Rajamani K., Allen-Ware M., Silva G., Heimsoth D., Ghose S., Mutlu O. 2018. CapMaestro: Exploiting Power Redundancy, Data Center-Wide Priorities, and Stranded Power for Boosting Data Center Performance, IBM Research Report, RC25680, March 28, 2018.
• Micron. 2017. Technical Note Calculating Memory Power for DDR4 SDRAM.
• Samsung (2016). Mobile DRAM Stack Specification.
• Ghose S., Yaliki A., Gupta R, et al. 2018.What Your DRAM Power Models Are Not Telling You: Lessons from a Detailed Experimental Study. SIGMETRICS 2018.
• Tanenbaum A. 2005. Structured Computer Organization (5th. ed.). Prentice Hall PTR,USA.
• Cheng W., Shen P., Li.X. 2019. Retention-Aware DRAM Auto-Refresh Scheme for Energyand Performance Efficiency Micromachines 2019, 10(9), 590.
• Gulur N.,Govindarajan, R., Mahesh M. 2016. MicroRefresh: Minimizing Refresh Overhead in DRAM Caches.MEMSYS '16: Proceedings of the Second InternationalSymposium on Memory Systems,pages 350–361.
• Mellor C. 2020, April 13. Why DRAM is stuck in a 10nm trap.Blocksandfiles.
• LaPedus M. 2016, February 18. 1xnm DRAM Challenges. Semi-engineering.
• MRAM-Info.2018-2020.
• RRAM-Info. 2018-2020.
• Subramanian L.,Usui H., Chang K., Mutlu O. 2016.DASH: Deadline-aware high-performance memory scheduler for heterogeneous systems with hardware accelerators. ACM Transactions on Architecture and Code Optimization (TACO).
• KayiranO. et al. 2019. Quantifying Data Locality in Dynamic Parallelism in GPUs. ACM SIGMETRICS Performance Evaluation Review 47(1):25-26.
• Tatar A., Krishnan R., Athanasopoulos E., Giuffrida C., Bos H., and Razavi K. 2018. Throwhammer: Rowhammer Attacks over the Network and Defenses. USENIX ATC ’18. July 11–13.
• Kim Y. et al. 2014. Flipping bits in memory without accessing them: An experimental study of DRAM disturbance errors. 2014 ACM/IEEE 41st International Symposium on Computer Architecture (ISCA), Minneapolis, MN, 2014, pp. 361-372.
• Seaborn M.2015, March 9. Exploiting the DRAM rowhammer bug to gain kernel privileges.Blog: Google Project Zero.
• Vusec. 2018.Drammer: Flip Feng Shui Goes Mobile.
• Google. 2015, August 10. Program for testing for the DRAM "rowhammer" problem. Github.
• Restifo M., Bernardi P., De Luca S., Sansonetti A. 2017. On-line software-based self-test for ECC of embedded RAM memories," 2017 IEEE International Symposium on Defect and Fault Tolerance in VLSI and Nanotechnology Systems (DFT), Cambridge, 2017, pp. 1-6.
• Newman L.H. 2018, November 21. An Ingenious Data Hack Is More Dangerous Than Anyone Feared.Wired.
• JEDEC. 2019, February 19. Updates Standard for Low Power Memory Devices: LPDDR5.