ABSTRACT
Recent studies reveal that one of the major challenges in scaling DRAM in deep sub-micron regime is its significant variations on cell restore time, which affects timing constraints such as write recovery time tWR. Adopting traditional approaches results in either low yield rate or large performance degradation. In this paper, we propose schemes to expose the variations to the architectural level. By constructing memory chunks with different accessing speeds and, in particular, exploiting the performance benefits of fast chunks, a variation-aware memory controller can effectively compensate the performance loss due to relaxed timing constraints. Our experimental results show that, comparing to traditional designs such as row sparing and ECC, the proposed schemes help to improve system performance by up to 10.3% and 12.9%, respectively, for 20nm and 14nm tech nodes on a 4-core multiprocessor system.
- "Co-architecting controllers and dram to enhance dram process scaling," in The Memory Forum, 2014.Google Scholar
- Samsung, "Strong 14nm finfet logic process and design infrastructure for advanced mobile soc applications," http://www.samsung.com/global/business/semiconductor/file/media/Samsung_Foundry_14nm_FinFET-0.pdf, 2013.Google Scholar
- R. Goering, "2014 tsmc technology symposium: Full speed ahead for 16nm finfet plus, 10nm, and 7nm," TSMC, Tech. Rep., April 2014.Google Scholar
- M. Mayberry, "Enabling breakthroughs in technology," Intel, Tech. Rep., June 2011.Google Scholar
- International Technology Roadmap for Semiconductors for Semiconductors, "Itrs report, 2012 edition," http://www.itrs.net.Google Scholar
- W. Mueller, G. Aichmayr, et al., "Challenges for the dram cell scaling to 40nm," in International Electron Devices Meeting, 2005.Google Scholar
- K. Kim, "Technology for sub-50nm dram and nand flash manufacturing," in International Electron Devices Meeting, 2005.Google Scholar
- T. Vogelsang, "Understanding the energy consumption of dynamic random access memories," in MICRO, 2010. Google ScholarDigital Library
- B. Jacob, S. Ng, et al., "Memory systems: Cache, dram, disk," in Morgan Kaufmann, 2007. Google ScholarDigital Library
- J. Liu, B. Jaiyen, et al., "Raidr: Retention-aware intelligent dram refresh," in ISCA, 2012. Google ScholarDigital Library
- A. Agrawal, A. Ansari, et al., "Mosaic: Exploiting the spatial locality of process variation to reduce refresh energy in on-chip edram modules," in ISCA, 2014.Google Scholar
- S. R. Sarangi, B. Greskamp, et al., "Varius: A model of process variation and resulting timing errors for microarchitects," IEEE Transactions on Semiconductor Manufacturing, 21(1): 3--13, 2008.Google ScholarCross Ref
- T. Karnik, S. Borkar, et al., "Probalistic and variation-tolerant design: Key to continued moore's law," in TAU Workshop, 2004.Google Scholar
- Samsung, "What is twr." {Online}. Available: http://www.samsung.com/global/business/semiconductor/file/product/tWR-0.pdfGoogle Scholar
- J. H. Ahn, N. P. Jouppi, et al., "Future scaling of processor-memory interfaces," in International Conference for High Performance Computing, Networking, Storage and Analysis, 2009. Google ScholarDigital Library
- H. Zheng, J. Lin, et al., "Mini-rank: Adaptive dram architecture for improving memory power efficiency," in MICRO, 2008. Google ScholarDigital Library
- "2gb ddr3 sdram data sheet," Hynix Semiconductor, 2010.Google Scholar
- T. Kirihata, Y. Watanabe, et al., "Fault-tolerant designs for 256 mb dram," IEEE Journal of Solid-State Circuits, 31(4): 558--566, 1996.Google ScholarCross Ref
- I. Koren and C. M. Krishna, Fault-Tolerant Systems. Morgan Kaufmann, 2010.Google ScholarDigital Library
- C.-L. Su, Y.-T. Yeh, et al., "An integrated ecc and redundancy repair scheme for memory reliability enhancement," in IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems, 2005. Google ScholarDigital Library
- Y. Kim, V. Seshadriet al., "A Case for Exploiting Subarray-Level Parallelism (SALP) in DRAM," in ISCA, 2012. Google ScholarDigital Library
- W. Kong, P. Parries, et al., "Analysis of Retention Time Distribution of Embedded DRAM - A New Method to Characterize Across-Chip Threshold Voltage Variation," in International Test Conference, 2008.Google Scholar
- K. Kim, J. Lee, "A New Investigation of Data Retention Time in Truly Nanoscaled DRAMs," in IEEE Electron Device Letters, 30(8): 846--848, 2009.Google ScholarCross Ref
- J. Kim, M. C. Papaefthymiou, "Block-based Multi-period Refresh for Energy Efficient Dynamic Memory," in IEEE International ASIC/SOC Conference, 2001.Google Scholar
- T. Ohsawa, K. Kai, et al., "Optimizaing the DRAM Refresh Count for Merged DRAM/Logilsis," in ISLPED, 1998. Google ScholarDigital Library
- R. K. Venkatesan, S. Herr, et al., "Retention-aware Placement in DRAM (rapid): Software Methods for Quasi-non-volatile DRAM," in HPCA, 2006.Google Scholar
Index Terms
- Exploiting DRAM restore time variations in deep sub-micron scaling
Recommendations
On the Restore Time Variations of Future DRAM Memory
Special Section of IDEA: Integrating Dataflow, Embedded Computing, and ArchitectureAs the de facto main memory standard, DRAM (Dynamic Random Access Memory) has achieved dramatic density improvement in the past four decades, along with the advancements in process technology. Recent studies reveal that one of the major challenges in ...
Exploiting Refresh Effect of DRAM Read Operations: A Practical Approach to Low-Power Refresh
Dynamic random access memory (DRAM) requires periodic refresh operations to retain its data. In practice, DRAM retention times are normally distributed from 64 ms to several seconds. However, the conventional refresh method uses 64 ms as the refresh ...
Selective restore: an energy efficient read disturbance mitigation scheme for future STT-MRAM
DAC '15: Proceedings of the 52nd Annual Design Automation ConferenceSTT-MRAM (Spin-Transfer Torque Magnetic RAM) has recently emerged as one of the most promising memory technologies for constructing large capacity last level cache (LLC) of low power mobile processors. With fast technology scaling, STT-MRAM read ...
Comments