ABSTRACT
The continuous rise of on-chip components like cores and caches has brought enormous computing capabilities at the cost of high leakage power and temperature. A recent study has shown a substantial spatial temperature variance in modern large on-chip caches. This high temperature elevates the cooling cost and becomes responsible for the thermal breakdown of the chip. One solution to reduce the leakage is the use of non-volatile memory (NVM) like STT-RAM. Other includes incorporating the concept of dark silicon. In this paper, we amalgamate the idea of using STT-RAM in the last level cache (LLC) and the dark silicon approach to shut down certain cache ways to leverage the benefits from both. We address the downsides like higher access latencies of STT-RAM by the use of hybrid cache (SRAM + STT-RAM) and weak endurance of the STT-RAM by wear leveling. We propose a system to handle three different temperature thresholds (high, medium, and low) by appropriately selecting the type of cache ways to be powered off. The proposed system delivers up to 5.38 K reduction in temperature compared to the baseline, 93% reduction in leakage power with an EDP gain up to 92%.
Supplemental Material
- Alessandro Bardine et al. 2007 a. Analysis of static and dynamic energy consumption in nuca caches: Initial results. In MEDEA. ACM, 105--112.Google Scholar
- Alankar V Umdekar et al. 2018. Dynamic Thermal Management by Using Task Migration in Conjunction with Frequency Scaling for Chip Multiprocessors. In VLSID. IEEE.Google Scholar
- Christian Bienia et al. 2008a. The PARSEC benchmark suite: Characterization and architectural implications. In PACT. ACM, 72--81.Google Scholar
- Dmytro Apalkov et al. 2013a. Spin-transfer torque magnetic random access memory (STT-MRAM). JETC, Vol. 9, 2 (2013), 13.Google Scholar
- Guangyu Sun et al. 2012a. Performance/thermal-aware design of 3D-stacked L2 caches for CMPs. TODAES, Vol. 17, 2 (2012), 13.Google Scholar
- Hadi Esmaeilzadeh et al. 2011a. Dark silicon and the end of multicore scaling. In ISCA. IEEE, 365--376.Google Scholar
- Ing-Chao Lin et al. 2015a. High-Endurance Hybrid Cache Design in CMP Architecture With Cache Partitioning and Access-Aware Policies. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, Vol. 23, 10 (2015), 2149--2161.Google ScholarCross Ref
- Ja Chun Ku et al. 2007 b. Thermal management of on-chip caches through power density minimization. TVLSI, Vol. 15, 5 (2007), 592--604.Google Scholar
- Joonho Kong et al. 2012b. Recent thermal management techniques for microprocessors. CSUR, Vol. 44, 3 (2012), 13.Google Scholar
- Jue Wang et al. 2013b. i 2 WAP: Improving non-volatile cache lifetime by reducing inter-and intra-set write variations. In HPCA. IEEE, 234--245.Google Scholar
- Kyriakos Stavrou et al. 2005. TSIC: Thermal Scheduling Simulator for Chip Multiprocessors, Vol. 3746. 589--599. https://doi.org/10.1007/11573036_56Google Scholar
- Mengquan Li et al. 2017. Chip temperature optimization for dark silicon many-core systems. TCAD, Vol. 37, 5 (2017), 941--953.Google Scholar
- Mesut Meterelliyoz et al. 2009a. Analysis of SRAM and eDRAM cache memories under spatial temperature variations. IEEE TCAD, Vol. 29, 1 (2009), 2--13.Google Scholar
- Muhammad Shafique et al. 2014a. Dark silicon as a challenge for hardware/software co-design: Invited special session paper. In CODES. ACM, 13.Google Scholar
- Muhammad Shafique et al. 2016. Computing in the dark silicon era: Current trends and research challenges. Design & Test, Vol. 34, 2 (2016), 8--23.Google ScholarCross Ref
- Nathan Binkert et al. 2011b. The gem5 simulator. SIGARCH, Vol. 39, 2 (2011), 1--7.Google ScholarDigital Library
- Raid Ayoub et al. 2010. Performance and energy efficient cache migrationapproach for thermal management in embedded systems. In GLSVLSI. ACM, 365--368.Google Scholar
- Runjie Zhang et al. 2015b. Hotspot 6.0: Validation, acceleration and extension. University of Virginia, Tech. Rep (2015).Google Scholar
- Shounak Chakraborty et al. 2019. Exploring the Role of Large Centralised Caches in Thermal Efficient Chip Design. ACM Trans. Des. Autom. Electron. Syst., Vol. 24, 5 (2019).Google Scholar
- S. Li et al. 2009b. McPAT: An integrated power, area, and timing modeling framework for multicore and manycore architectures. In MICRO. 469--480.Google Scholar
- Sparsh Mittal et al. 2014b. LastingNVCache: A technique for improving the lifetime of non-volatile caches. In ISVLSI. IEEE, 534--540.Google Scholar
- Sobhan Niknam et al. 2015c. Energy efficient 3D Hybrid processor-memory architecture for the dark silicon age. In ReCoSoC. IEEE, 1--8.Google Scholar
- Shyamkumar Thoziyoor et al. 2008b. CACTI 5.1. Technical Report. Technical Report, HP Labs.Google Scholar
- V. Hanumaiah et al. 2009c. Maximizing performance of thermally constrained multi-core processors by dynamic voltage and frequency control. In ICCAD. 310--313.Google Scholar
- Wei Zang et al. 2013c. A survey on cache tuning from a power/energy perspective. CSUR, Vol. 45, 3 (2013), 32.Google Scholar
- Xiangyu Dong et al. 2012c. NVSim: A Circuit-Level Performance, Energy, and Area Model for Emerging Nonvolatile Memory. TCAD, Vol. 31 (2012), 994--1007.Google Scholar
- Xiaoxia Wu et al. 2009 d. Hybrid cache architecture with disparate memory technologies. In SIGARCH, Vol. 37. ACM, 34--45.Google Scholar
- Yu-Ting Chen et al. 2012 d. Dynamically reconfigurable hybrid cache: An energy-efficient last-level cache design. In DATE. EDA Consortium, 45--50.Google Scholar
Index Terms
- Dimming Hybrid Caches to Assist in Temperature Control of Chip MultiProcessors
Recommendations
Performance linked dynamic cache tuning
Advancement in semiconductor technology increases power density in recent Chip Multi-Processors (CMPs) which significantly increases the leakage energy consumptions of on-chip Last Level Caches (LLCs). Performance linked dynamic tuning in LLC size is a ...
Statistical Peak Temperature Prediction and Thermal Yield Improvement for 3D Chip Multiprocessors
Thermal issues have become critical roadblocks for achieving highly reliable three-dimensional (3D) integrated circuits (ICs). The presence of process variations further exacerbates these problems. In this article, we propose techniques for the ...
Exploring the Role of Large Centralised Caches in Thermal Efficient Chip Design
In the era of short channel length, Dynamic Thermal Management (DTM) has become a challenging task for the architects and designers engineering modern Chip Multi-Processors (CMPs). Ever-increasing demand of processing power along with the developed ...
Comments