pith. machine review for the scientific record. sign in

arxiv: 2604.19932 · v1 · submitted 2026-04-21 · 💻 cs.AR

Recognition: unknown

Efficient Page Migration in Hybrid Memory Systems

Upasna, Venkata Kalyan Tavva

Authors on Pith no claims yet

Pith reviewed 2026-05-10 01:10 UTC · model grok-4.3

classification 💻 cs.AR
keywords page migrationhybrid memoryheterogeneous memoryTLBpage tableflat address spaceIPC
0
0 comments X

The pith

Duon stores updated page mappings directly in the TLB and page table to avoid shootdowns and invalidations during migration in hybrid memory systems.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes Duon to move frequently used pages into faster memory without the usual costs in flat-address hybrid systems. Normally such moves force TLB shootdowns and cache invalidations that slow the processor. Duon places the new mapping straight into the TLB and page table so those steps are skipped. The method works with any existing migration policy. Measured results show a 3.87 percent rise in instructions per cycle over earlier techniques.

Core claim

In a flat address space that pools high-bandwidth memory with slower DRAM or NVM, page migration to faster tiers normally requires TLB shootdowns and cache line invalidations. Duon eliminates these steps by writing the updated mapping information directly into the TLB and page table for the remapped pages.

What carries the argument

Duon, a mechanism that stores the new mapping for each migrated page directly inside the TLB and page table entries so that shootdowns and invalidations are no longer needed.

If this is right

  • Page migration no longer triggers TLB shootdowns.
  • Cache lines remain valid after a page is relocated.
  • Any existing page migration policy can be used without added overhead.
  • Overall instructions-per-cycle performance rises by 3.87 percent compared with prior methods.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • More frequent page moves become practical because their cost drops.
  • The same direct-update idea could apply to other large-memory remapping schemes.
  • Lower migration overhead may allow systems to keep more data in fast memory and reduce energy spent on slower tiers.

Load-bearing premise

Directly writing new mappings into the TLB and page table after migration keeps the system correct and does not create hidden performance or coherence problems on real hardware.

What would settle it

A workload run with Duon that still shows accesses using stale TLB entries or cache coherence errors after a page has been moved.

Figures

Figures reproduced from arXiv: 2604.19932 by Upasna, Venkata Kalyan Tavva.

Figure 1
Figure 1. Figure 1: Overview of two design choices of Heterogeneous Memory. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Accumulated overhead cycles per core in ONFLY and EPOCH. Y-axis is in logarithmic scale. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Cache and TLB overhead cycles per epoch in EPOCH. Y-axis is in logarithmic scale. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Extended Page Table and TLB Structure. (Note: Column headers shaded in gray represent existing [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Page migration demonstration in Unified Address Space. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Overview of steps involved in Page Migration. [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Remapped Physical Address updation in Extended TLB and Page Table. [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Extended TLB and Page Table Lookup in Duon. to migrate, identifying the victim page in the target memory, etc., are all done by the migration controller. The migration controller makes sure that the actual migration of the pages happens and updates its fields from time to time during the migration in TLB and EPT. Hence, Duon can be integrated into any of the existing page migration policies such as ONFLY[9… view at source ↗
Figure 9
Figure 9. Figure 9: Normalized IPC improvement of various HMA techniques, for threshold value of 64, HBM Size: 1 GB, [PITH_FULL_IMAGE:figures/full_fig_p017_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: (a) Normalized IPC is reported for ONFLY-DUON when compared with ONFLY, EPOCH-DUON [PITH_FULL_IMAGE:figures/full_fig_p017_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Normalized IPC improvement with DUON for ONFLY and EPOCH, for threshold values of 64 and [PITH_FULL_IMAGE:figures/full_fig_p019_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Normalized IPC improvement with DUON for ONFLY and EPOCH, for threshold values of 64 and [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Normalized IPC improvement of ONFLY and EPOCH using DUON, with threshold value of 128,for [PITH_FULL_IMAGE:figures/full_fig_p020_13.png] view at source ↗
read the original abstract

Heterogeneous Memory Architecture (HMA) aims to optimize memory usage by leveraging a combination of memory types, such as high-bandwidth memory (HBM), commodity DRAM, and non-volatile memory (NVM), when utilized as main memory. To achieve maximum performance benefits, frequently accessed data pages are prioritized for storage in the faster HBM, while less frequently accessed pages are stored in slower memory types like DRAM or NVM. This enables a more efficient allocation of memory resources and improves overall system performance. In a Flat Address Space memory organization, all memory types, both fast and slow, are treated as a unified memory pool. This approach increases the overall memory capacity accessible to the system. In Flat Address Space organization, frequently accessed data pages may need to be remapped from slower memory to faster memory to improve memory access times. Such relocation requires changes to the data/states in the TLB (TLB shootdown) and the processor cache (cache line invalidations), leading to performance degradation. To address these inefficiencies, we propose a novel solution called Duon. The goal of Duon is to eliminate the overheads associated with page migration in systems using Extended TLB and Page Table. Specifically, our approach ensures that the updated mapping information for remapped pages is carefully stored directly in the TLB and page table itself. By doing so, the need for TLB shootdown and cache line invalidation after page migration is eliminated. Consequently, our proposal results in an overall improvement in IPC by 3.87% over existing state-of-the-art techniques, enhancing the efficiency and performance of heterogeneous memory systems. Further, our approach can work with any of the existing page migration policies and improve the performance.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript proposes Duon, a technique for efficient page migration in heterogeneous memory systems (HMA) organized as a flat address space. It argues that by directly storing updated page mappings in an extended TLB and page table during migration from slower to faster memory (e.g., DRAM/NVM to HBM), the overheads of TLB shootdowns and cache-line invalidations are eliminated. The approach is claimed to be compatible with any existing migration policy and yields a 3.87% IPC improvement over state-of-the-art methods.

Significance. If the coherence mechanism is sound and the performance gain is reproducible, the work could reduce migration costs in hybrid memory architectures, benefiting systems that combine HBM with commodity DRAM or NVM. The compatibility claim with existing policies is a potential strength, but the absence of any correctness argument, pseudocode, or hardware model in the abstract leaves the significance highly conditional.

major comments (2)
  1. [Abstract] Abstract: The central claim that 'carefully storing' updated mappings directly in the TLB and page table eliminates TLB shootdown and cache-line invalidation is load-bearing for the 3.87% IPC result, yet no mechanism, coherence protocol, or hardware assumption is provided. In standard multi-core x86/ARM systems, a page-table write on one core leaves stale TLB entries on others unless an IPI-based shootdown (or equivalent) is issued; the manuscript must specify how Duon preserves translation coherence without these steps or without non-standard hardware support.
  2. [Abstract] Abstract: The performance claim of a 3.87% IPC improvement over state-of-the-art techniques is presented without any reference to evaluation methodology, benchmarks, workloads, simulation parameters, or baseline implementations. This makes it impossible to assess whether the gain is attributable to the proposed coherence elimination or to other unstated factors.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for your review and constructive feedback on our manuscript. We appreciate the points raised about the abstract and will revise it to provide additional clarity on the mechanism and evaluation details while preserving the overall contribution. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that 'carefully storing' updated mappings directly in the TLB and page table eliminates TLB shootdown and cache-line invalidation is load-bearing for the 3.87% IPC result, yet no mechanism, coherence protocol, or hardware assumption is provided. In standard multi-core x86/ARM systems, a page-table write on one core leaves stale TLB entries on others unless an IPI-based shootdown (or equivalent) is issued; the manuscript must specify how Duon preserves translation coherence without these steps or without non-standard hardware support.

    Authors: We agree that the abstract is too concise to convey the coherence mechanism. The full manuscript describes Duon as relying on an extended TLB design that performs in-place atomic updates to page mappings, with hardware-level propagation ensuring all cores observe the new translation without software shootdowns or invalidations. This assumes an extended TLB supporting direct coherence for mapping changes, as outlined in the design section. To address the concern, we will revise the abstract to briefly state the hardware assumptions and reference the detailed protocol explanation (including any pseudocode) in the body. We will also expand the relevant sections if needed to strengthen the correctness argument. revision: yes

  2. Referee: [Abstract] Abstract: The performance claim of a 3.87% IPC improvement over state-of-the-art techniques is presented without any reference to evaluation methodology, benchmarks, workloads, simulation parameters, or baseline implementations. This makes it impossible to assess whether the gain is attributable to the proposed coherence elimination or to other unstated factors.

    Authors: We agree the abstract omits evaluation context due to length limits. The manuscript's evaluation section details the methodology, including cycle-accurate simulation, benchmarks, and baselines used to obtain the 3.87% IPC gain. We will revise the abstract to include a brief reference to the evaluation setup (e.g., simulation framework and workload characteristics) so readers can better attribute the reported improvement. revision: yes

Circularity Check

0 steps flagged

No circularity; engineering proposal with no derivation chain or fitted inputs

full rationale

The manuscript proposes Duon, a technique to avoid TLB shootdowns and cache invalidations during page migration by directly updating mappings in the TLB and page table. The abstract and provided text contain no equations, parameters fitted to data, self-citations used as load-bearing premises, uniqueness theorems, or ansatzes. The 3.87% IPC claim is presented as an empirical outcome of the proposal rather than a mathematical reduction to prior results. No step reduces by construction to its own inputs; the work is a self-contained systems design evaluated against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based solely on the abstract, the proposal rests on standard assumptions about memory hierarchy behavior and hardware support for extended TLBs; no free parameters, new axioms, or invented entities are explicitly introduced or quantified.

pith-pipeline@v0.9.0 · 5606 in / 1169 out tokens · 54220 ms · 2026-05-10T01:10:17.063213+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 33 canonical work pages

  1. [1]

    Shashank Adavally, Mahzabeen Islam, and Krishna Kavi. 2021. Dynamically Adapting Page Migration Policies Based on Applications’ Memory Access Behaviors. J. Emerg. Technol. Comput. Syst. 17, 2, Article 16 (March 2021), 24 pages. doi:10.1145/3444750 , Vol. 1, No. 1, Article . Publication date: April 2026. Efficient Page Migration in Hybrid Memory Systems 21

    work page doi:10.1145/3444750 2021

  2. [2]

    Shashank Adavally and Shashankadavally. 2021. Subpage Migration in Heterogeneous Memory Systems. https: //api.semanticscholar.org/CorpusID:252564184

    2021

  3. [3]

    Christian Bienia, Sanjeev Kumar, Jaswinder Pal Singh, and Kai Li. 2008. The PARSEC benchmark suite: Characterization and architectural implications. In2008 International Conference on Parallel Architectures and Compilation Techniques (PACT). 72–81

    2008

  4. [4]

    E. Chen, D. Lottis, A. Driskill-Smith, D. Druist, V. Nikitin, S. Watts, X. Tang, and D. Apalkov. 2010. Non-volatile spin-transfer torque RAM (STT-RAM). In 68th Device Research Conference. 249–252. doi:10.1109/DRC.2010.5551975

    work page doi:10.1109/drc.2010.5551975 2010

  5. [5]

    Chiachen Chou, Aamer Jaleel, and Moinuddin Qureshi. 2017. BATMAN: techniques for maximizing system bandwidth of memory systems with stacked-DRAM. In Proceedings of the International Symposium on Memory Systems (Alexandria, Virginia) (MEMSYS ’17). Association for Computing Machinery, New York, NY, USA, 268–280. doi:10.1145/3132402.3132404

    work page doi:10.1145/3132402.3132404 2017

  6. [6]

    Chia Chen Chou, Aamer Jaleel, and Moinuddin K. Qureshi. 2014. CAMEO: A Two-Level Memory Organization with Capacity of Main Memory and Flexibility of Hardware-Managed Cache. In 2014 47th Annual IEEE/ACM International Symposium on Microarchitecture. 1–12. doi:10.1109/MICRO.2014.63

    work page doi:10.1109/micro.2014.63 2014

  7. [7]

    Penglin Gao, Zhaoming Han, and Fucheng Wan. 2020. Big Data Processing and Application Research. In 2020 2nd International Conference on Artificial Intelligence and Advanced Manufacture (AIAM). 125–128. doi:10.1109/ AIAM50918.2020.00031

    work page arXiv 2020

  8. [8]

    Yuncheng Guo, Yu Hua, and Pengfei Zuo. 2018. DFPC: A dynamic frequent pattern compression scheme in NVM- based main memory. In 2018 Design, Automation & Test in Europe Conference & Exhibition (DATE). 1622–1627. doi:10.23919/DATE.2018.8342274

    work page doi:10.23919/date.2018.8342274 2018

  9. [9]

    Mahzabeen Islam, Shashank Adavally, Marko Scrbak, and Krishna Kavi. 2020. On-the-fly Page Migration and Address Reconciliation for Heterogeneous Memory Systems. J. Emerg. Technol. Comput. Syst. 16, 1, Article 10 (Jan. 2020), 27 pages. doi:10.1145/3364179

    work page doi:10.1145/3364179 2020

  10. [10]

    Joe Jeddeloh and Brent Keeth. 2012. Hybrid memory cube new DRAM architecture increases density and performance. In 2012 Symposium on VLSI Technology (VLSIT). 87–88. doi:10.1109/VLSIT.2012.6242474

    work page doi:10.1109/vlsit.2012.6242474 2012

  11. [11]

    JEDEC. 2023. Low Power Double Data Rate (LPDDR) 5/5X. https://www.jedec.org/document_search?search_api_ views_fulltext=jesd209

    2023

  12. [12]

    Loh, Cansu Kaynak, and Babak Falsafi

    Djordje Jevdjic, Gabriel H. Loh, Cansu Kaynak, and Babak Falsafi. 2014. Unison Cache: A Scalable and Effective Die- Stacked DRAM Cache. In Proceedings of the 47th Annual IEEE/ACM International Symposium on Microarchitecture (Cambridge, United Kingdom) (MICRO-47). IEEE Computer Society, USA, 25–37. doi:10.1109/MICRO.2014.51

    work page doi:10.1109/micro.2014.51 2014

  13. [13]

    Djordje Jevdjic, Stavros Volos, and Babak Falsafi. 2013. Die-stacked DRAM caches for servers: hit ratio, latency, or bandwidth? have it all with footprint cache. SIGARCH Comput. Archit. News 41, 3 (June 2013), 404–415. doi:10.1145/ 2508148.2485957

    work page arXiv 2013

  14. [14]

    Hongshin Jun, Jinhee Cho, Kangseol Lee, Ho-Young Son, Kwiwook Kim, Hanho Jin, and Keith Kim. 2017. HBM (High Bandwidth Memory) DRAM Technology and Architecture. In2017 IEEE International Memory Workshop (IMW). 1–4. doi:10.1109/IMW.2017.7939084

    work page doi:10.1109/imw.2017.7939084 2017

  15. [15]

    Shivanjali Khare and Michael Totaro. 2019. Big Data in IoT. In 2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT). 1–7. doi:10.1109/ICCCNT45670.2019.8944495

    work page doi:10.1109/icccnt45670.2019.8944495 2019

  16. [16]

    Jung-Sik Kim, Chi Sung Oh, Hocheol Lee, Donghyuk Lee, Hyong Ryol Hwang, Sooman Hwang, Byongwook Na, Joungwook Moon, Jin-Guk Kim, Hanna Park, Jang-Woo Ryu, Kiwon Park, Sang Kyu Kang, So-Young Kim, Hoyoung Kim, Jong-Min Bang, Hyunyoon Cho, Minsoo Jang, Cheolmin Han, Jung-Bae LeeLee, Joo Sun Choi, and Young-Hyun Jun. 2012. A 1.2 V 12.8 GB/s 2 Gb Mobile Wide-...

    work page doi:10.1109/jssc.2011.2164731 2012

  17. [17]

    Youngin Kim, Hyeonjin Kim, and William J. Song. 2023. NOMAD: Enabling Non-blocking OS-managed DRAM Cache via Tag-Data Decoupling. In 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA). 193–205. doi:10.1109/HPCA56546.2023.10071016

    work page doi:10.1109/hpca56546.2023.10071016 2023

  18. [18]

    Yoongu Kim, Weikun Yang, and Onur Mutlu. 2016. Ramulator: A Fast and Extensible DRAM Simulator.IEEE Computer Architecture Letters 15, 1 (2016), 45–49. doi:10.1109/LCA.2015.2414456

    work page doi:10.1109/lca.2015.2414456 2016

  19. [19]

    Apostolos Kokolis, Dimitrios Skarlatos, and Josep Torrellas. 2019. PageSeer: Using Page Walks to Trigger Page Swaps in Hybrid Memory Systems. In 2019 IEEE International Symposium on High Performance Computer Architecture (HPCA). 596–608. doi:10.1109/HPCA.2019.00012

    work page doi:10.1109/hpca.2019.00012 2019

  20. [20]

    Harnessing GPU tensor cores for fast FP16 arithmetic to speed up mixed- precision iterative refinement solvers,

    Jagadish B. Kotra, Haibo Zhang, Alaa R. Alameldeen, Chris Wilkerson, and Mahmut T. Kan- demir. 2018. CHAMELEON: A Dynamically reconfigurable heterogeneous memory system. In Proceedings - 51st Annual IEEE/ACM International Symposium on Microarchitecture, MICRO 2018 (Proceedings of the Annual International Symposium on Microarchitecture, MICRO). IEEE Comput...

    work page doi:10.1109/micro.2018.00050 2018

  21. [21]

    Yongjun Lee, Jongwon Kim, Hakbeom Jang, Hyunggyun Yang, Jangwoo Kim, Jinkyu Jeong, and Jae W. Lee. 2015. A fully associative, tagless DRAM cache. In Proceedings of the 42nd Annual International Symposium on Computer Architecture (Portland, Oregon) (ISCA ’15). Association for Computing Machinery, New York, NY, USA, 211–222. doi:10.1145/2749469.2750383

    work page doi:10.1145/2749469.2750383 2015

  22. [22]

    Shengmei Li, Buqi Cheng, Xingyu Gao, Lin Qiao, and Zhizhong Tang. 2009. Performance Characterization of SPEC CPU2006 Benchmarks on Intel and AMD Platform. In 2009 First International Workshop on Education Technology and Computer Science, Vol. 2. 116–121. doi:10.1109/ETCS.2009.288

    work page doi:10.1109/etcs.2009.288 2009

  23. [23]

    Haikun Liu, Yujie Chen, Xiaofei Liao, Hai Jin, Bingsheng He, Long Zheng, and Rentong Guo. 2017. Hardware/software cooperative caching for hybrid DRAM/NVM memory architectures. In Proceedings of the International Conference on Supercomputing (Chicago, Illinois) (ICS ’17). Association for Computing Machinery, New York, NY, USA, Article 26, 10 pages. doi:10....

    work page doi:10.1145/3079079.3079089 2017

  24. [24]

    Jihang Liu and Shimin Chen. 2019. Initial Experience with 3D XPoint Main Memory. In 2019 IEEE 35th International Conference on Data Engineering Workshops (ICDEW). 300–305. doi:10.1109/ICDEW.2019.00009

    work page doi:10.1109/icdew.2019.00009 2019

  25. [25]

    Loh and Mark D

    Gabriel H. Loh and Mark D. Hill. 2011. Efficiently enabling conventional block sizes for very large die-stacked DRAM caches. In 2011 44th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO). 454–464

    2011

  26. [26]

    Meswani, Sergey Blagodurov, David Roberts, John Slice, Mike Ignatowski, and Gabriel H

    Mitesh R. Meswani, Sergey Blagodurov, David Roberts, John Slice, Mike Ignatowski, and Gabriel H. Loh. 2015. Heterogeneous memory architectures: A HW/SW approach for mixing die-stacked and off-package memories. In 2015 IEEE 21st International Symposium on High Performance Computer Architecture (HPCA). 126–136. doi:10.1109/ HPCA.2015.7056027

    work page arXiv 2015

  27. [27]

    OpenAI. 2022. ChatGPT. https://chat.openai.com

    2022

  28. [28]

    Qureshi and Gabe H

    Moinuddin K. Qureshi and Gabe H. Loh. 2012. Fundamental Latency Trade-off in Architecting DRAM Caches: Outper- forming Impractical SRAM-Tags with a Simple and Practical Design. In 2012 45th Annual IEEE/ACM International Symposium on Microarchitecture. 235–246. doi:10.1109/MICRO.2012.30

    work page doi:10.1109/micro.2012.30 2012

  29. [29]

    David Patterson Scott Beamer, Krste Asanović. 2015. The GAP Benchmark Suite. https://arxiv.org/abs/1508.03619

    work page arXiv 2015

  30. [30]

    Shihao Song, Anup Das, Onur Mutlu, and Nagarajan Kandasamy. 2020. Improving phase change memory performance with data content aware access. In Proceedings of the 2020 ACM SIGPLAN International Symposium on Memory Management (London, UK) (ISMM 2020). Association for Computing Machinery, New York, NY, USA, 30–47. doi:10. 1145/3381898.3397210

    work page arXiv 2020

  31. [31]

    Arun Subramaniyan, Yufeng Gu, Timothy Dunn, Somnath Paul, Md Vasimuddin, Sanchit Misra, David Blaauw, Satish Narayanasamy, and Reetuparna Das. 2021. GenomicsBench: A Benchmark Suite for Genomics. In 2021 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS). 1–12. doi:10.1109/ISPASS51385. 2021.00012

    work page doi:10.1109/ispass51385 2021

  32. [32]

    Nittala Swapna Suhasini and Srilatha Puli. 2021. Big Data Analytics in Cloud Computing. In 2021 Sixth International Conference on Image Information Processing (ICIIP), Vol. 6. 320–325. doi:10.1109/ICIIP53038.2021.9702705

    work page doi:10.1109/iciip53038.2021.9702705 2021

  33. [33]

    Evangelos Vasilakis, Vassilis Papaefstathiou, Pedro Trancoso, and Ioannis Sourdis. 2020. Hybrid2: Combining Caching and Migration in Hybrid Memory Systems. In 2020 IEEE International Symposium on High Performance Computer Architecture (HPCA). 649–662. doi:10.1109/HPCA47549.2020.00059

    work page doi:10.1109/hpca47549.2020.00059 2020

  34. [34]

    R Matthew Ward, Robert Schmieder, Gareth Highnam, and David Mittelman and. 2013. Big data challenges and opportunities in high-throughput sequencing. Systems Biomedicine 1, 1 (2013), 29–34. doi:10.4161/sysb.24470 arXiv:https://doi.org/10.4161/sysb.24470

    work page doi:10.4161/sysb.24470 2013

  35. [35]

    Philip Wong, Simone Raoux, SangBum Kim, Jiale Liang, John P

    H.-S. Philip Wong, Simone Raoux, SangBum Kim, Jiale Liang, John P. Reifenberg, Bipin Rajendran, Mehdi Asheghi, and Kenneth E. Goodson. 2010. Phase Change Memory.Proc. IEEE 98, 12 (2010), 2201–2227. doi:10.1109/JPROC.2010.2070050

    work page doi:10.1109/jproc.2010.2070050 2010

  36. [36]

    Yinglong Xia, Ilie Gabriel Tanase, Lifeng Nai, Wei Tan, Yanbin Liu, Jason Crawford, and Ching-Yung Lin. 2014. Graph analytics and storage. In 2014 IEEE International Conference on Big Data (Big Data). 942–951. doi:10.1109/BigData. 2014.7004326

    work page doi:10.1109/bigdata 2014

  37. [37]

    Vinson Young, Chiachen Chou, Aamer Jaleel, and Moinuddin Qureshi. 2018. ACCORD: Enabling Associativity for Gigascale DRAM Caches by Coordinating Way-Install and Way-Prediction. In 2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA). 328–339. doi:10.1109/ISCA.2018.00036

    work page doi:10.1109/isca.2018.00036 2018

  38. [38]

    Hughes, Nadathur Satish, Onur Mutlu, and Srinivas Devadas

    Xiangyao Yu, Christopher J. Hughes, Nadathur Satish, Onur Mutlu, and Srinivas Devadas. 2017. Banshee: bandwidth-efficient DRAM caching via software/hardware cooperation. InProceedings of the 50th Annual IEEE/ACM International Symposium on Microarchitecture (Cambridge, Massachusetts) (MICRO-50 ’17). Association for Comput- ing Machinery, New York, NY, USA,...

    work page doi:10.1145/3123939.3124555 2017