pith. sign in

arxiv: 1907.02622 · v1 · pith:IM5QRNHBnew · submitted 2019-07-04 · 💻 cs.IT · eess.SP· math.IT

Low-power and Reliable Solid-state Drive with Inverted Limited Weight Coding

Armin Ahmadzadeh , Omid Hajihassani , Pooria Taheri , Seyed Hossein Khasteh This is my paper

Pith reviewed 2026-05-25 08:36 UTC · model grok-4.3

classification 💻 cs.IT eess.SPmath.IT
keywords Inverted Limited Weight CodingNAND flashsolid-state driveprogram energydata retentionthreshold voltagecell error ratecoding scheme
0
0 comments X

The pith

Inverted Limited Weight Coding uses flag bits to favor more 1s in NAND codewords, lowering threshold voltages to reduce SSD program energy and improve retention.

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

The paper introduces Inverted Limited Weight Coding to encode data for NAND flash memory so that more cells receive lower threshold voltages. It does this by preferring logical 1s and adding flag bits to mark the inversions. A reader would care because cells at lower voltages require less energy to program and hold their charge longer with less interference between neighbors. Tests on three SSDs show concrete drops in energy use, gains in how long data lasts, and fewer errors in multi-level cells.

Core claim

ILWC favors a greater number of 1s in generated codewords at the cost of flag-bit redundancy. This increase in 1s raises the number of cells with lower threshold voltages. Through those lower voltages, ILWC reduces SSD program energy consumption, increases data retention by lowering the intrinsic electric field, diminishes cell-to-cell coupling noise via a 35% reduction in worst-case threshold voltage shift, and lowers MLC cell error rates. The scheme achieves more than 20% reduction in program energy, more than 18% improvement in retention, 5.3% lower error rate, and 37.5% faster program performance.

What carries the argument

Inverted Limited Weight Coding (ILWC), which inverts selected bits to produce codewords containing more logical 1s, tracked by added flag bits, so that more NAND cells operate at lower threshold voltages.

If this is right

  • SSD program operation energy consumption drops by more than 20%.
  • Cells retain data longer because the intrinsic electric field decreases by more than 18%.
  • Worst-case threshold voltage shift in adjacent cells falls by 35%, cutting coupling noise.
  • MLC cell error rate decreases by 5.3%.
  • Program operation performance improves by 37.5%.

Where Pith is reading between the lines

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

  • The flag-bit overhead may become proportionally smaller on larger block sizes, increasing net savings in high-capacity drives.
  • Lower average threshold voltages could permit reduced program voltage settings in future NAND generations without raising error rates.
  • The same preference for 1s might combine with existing wear-leveling routines to slow cell degradation over the drive lifetime.

Load-bearing premise

The increase in logical 1s produced by the flag-bit mechanism will reliably translate into lower threshold voltages across real NAND cells and the added redundancy will not offset the reported energy and reliability gains.

What would settle it

Side-by-side measurement of actual program energy, threshold voltage distributions, and retention time on physical NAND chips running ILWC versus standard encoding under identical write workloads.

Figures

Figures reproduced from arXiv: 1907.02622 by Armin Ahmadzadeh, Omid Hajihassani, Pooria Taheri, Seyed Hossein Khasteh.

Figure 2
Figure 2. Figure 2: Threshold voltage levels of the 2-bit multi-level cell [5] The number of the bits representable by a 2-bit MLC FGT is twofold of the number of the bits that a single SLC cell can [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The internal architecture of a NAND-based solid-state drive [5] B. Limited Wight Coding Limited Weight Coding was proposed as an approach to decrease the I/O power dissipation in the bus lines [15]. As I/O transitions contribute directly to higher noise and more energy consumption, the proposed coding scheme reduces the number of transitions in data in order to reduce the energy consumption of the unit. Th… view at source ↗
Figure 4
Figure 4. Figure 4: The total Percentage of ‘1’s and overhead of different segment sizes in the perfect ILWC B. High Reliability The bit errors in the data that are stored in the solid-state drives are due to the SSD’s internal failure mechanisms, which have complex physical and architectural aspects [5]. We study the effect of the ILWC technique on data retention (detrapping), which is a physical aspect of reliability and re… view at source ↗
Figure 9
Figure 9. Figure 9: Probability of ‘1’s in 616 MP3 files achieved by uncoded data [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 17
Figure 17. Figure 17: MLC-A program energy consumption of different coding settings From these evaluations, we see that in SLC-A after applying the 2-ILWC coding to the codewords, the energy consumption drops by 17.75%. The more economical, in terms of better coding gain (see [PITH_FULL_IMAGE:figures/full_fig_p010_17.png] view at source ↗
Figure 15
Figure 15. Figure 15: SLC-A program energy consumption of different coding settings [PITH_FULL_IMAGE:figures/full_fig_p010_15.png] view at source ↗
read the original abstract

In this work, we propose a novel coding scheme which based on the characteristics of NAND flash cells, generates codewords that reduce the energy consumption and improve the reliability of solid-state drives. This novel coding scheme, namely Inverted Limited Weight Coding (ILWC), favors a greater number of '1's appearing in its generated codewords at the cost of added information redundancy, as a form of flag bits. This increase in the number of bits valued as logical '1', in the generated codewords, will increase the number of cells that have lower threshold voltages. Through cells with lower threshold voltages, ILWC fruitfully reduces the SSD's program operation energy consumption. Moreover, it increases the SSD's data retention rate and reliability by decreasing the threshold voltage of the cells. The evaluation of our proposed coding method on three different SSDs, indicates more than 20% reduction in the SSD's program operation energy consumption. In addition, ILWC improves the cells' data retention rate by decreasing their intrinsic electric field by more than 18%. Moreover, the SSD's cell-to-cell coupling noise is diminished with the help of 35% reduction in the worst-case threshold voltage shift in a cell's adjacent cells. All this leads to 5.3% reduction in the MLC's cell error rate. In addition, ILWC achieves 37.5% improvement in the performance of the SSD program operation.

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 / 2 minor

Summary. The manuscript proposes Inverted Limited Weight Coding (ILWC), a scheme that appends flag bits to invert selected data blocks and thereby increase the density of logical 1s in the written codewords. The higher 1-density is claimed to produce lower threshold voltages in NAND cells, which in turn yields >20% reduction in SSD program energy, >18% improvement in data retention, 35% reduction in worst-case threshold-voltage shift, 5.3% reduction in MLC cell error rate, and 37.5% improvement in program performance. These figures are obtained from hardware measurements on three SSDs.

Significance. If the reported gains can be shown to survive after subtracting the redundancy overhead of the flag bits, the technique would constitute a low-overhead, software-only method for improving both energy and reliability of existing NAND-based SSDs. The empirical evaluation on real hardware is a positive feature; however, the absence of protocol details, baselines, and net-gain accounting currently prevents independent verification of the headline numbers.

major comments (2)
  1. [Abstract] Abstract: the central quantitative claims (>20% energy reduction, >18% retention improvement, 5.3% error-rate reduction, etc.) are presented as aggregate percentages without an explicit net-energy or net-reliability formula that subtracts the cost of the added flag bits. Because ILWC necessarily writes extra cells, the manuscript must demonstrate that the higher 1-density still produces net gains; no gross-vs-net breakdown or overhead accounting appears in the reported SSD results.
  2. [Abstract] Abstract (and Evaluation section): the experimental description supplies neither the coding baselines used for comparison, the precise SSD models and configurations, the number of trials, nor error bars. Without these elements the reported percentages cannot be reproduced or assessed for statistical significance, undermining the load-bearing empirical support for the performance claims.
minor comments (2)
  1. [Abstract] The abstract states that ILWC “favors a greater number of ‘1’s … at the cost of added information redundancy, as a form of flag bits,” yet never quantifies the typical flag-bit overhead (bits per block or percentage).
  2. Notation for the flag-bit mechanism and the inversion decision rule is introduced only informally; a compact algorithmic description or pseudocode would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on clarifying net gains after overhead and improving experimental reproducibility. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central quantitative claims (>20% energy reduction, >18% retention improvement, 5.3% error-rate reduction, etc.) are presented as aggregate percentages without an explicit net-energy or net-reliability formula that subtracts the cost of the added flag bits. Because ILWC necessarily writes extra cells, the manuscript must demonstrate that the higher 1-density still produces net gains; no gross-vs-net breakdown or overhead accounting appears in the reported SSD results.

    Authors: The reported figures were measured directly on three physical SSDs running the full ILWC implementation, which includes writing the flag bits. The hardware results therefore already incorporate the redundancy overhead and reflect net gains. To make this explicit, we will add a dedicated paragraph in the revised manuscript that quantifies the flag-bit overhead and confirms that all SSD measurements include the complete encoding process. revision: yes

  2. Referee: [Abstract] Abstract (and Evaluation section): the experimental description supplies neither the coding baselines used for comparison, the precise SSD models and configurations, the number of trials, nor error bars. Without these elements the reported percentages cannot be reproduced or assessed for statistical significance, undermining the load-bearing empirical support for the performance claims.

    Authors: We agree that the current description lacks sufficient detail for independent reproduction. In the revised manuscript we will expand the Evaluation section to name the exact SSD models and configurations, state the coding baselines used for comparison, report the number of trials performed, and include error bars or other statistical measures. revision: yes

Circularity Check

0 steps flagged

No circularity; claims rest on empirical SSD evaluations

full rationale

The manuscript presents ILWC as a coding scheme that favors more 1s via flag-bit redundancy and then reports measured gains (energy, retention, error rate) from running the scheme on three SSDs. No derivation chain, equations, fitted parameters renamed as predictions, or self-citation load-bearing steps appear in the text. The reported percentages are outputs of the evaluation rather than quantities forced by construction from the scheme definition itself.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The central claim rests on the domain assumption that NAND cells exhibit lower threshold voltage for logical 1 than for 0 and that this difference directly controls program energy and retention; no explicit free parameters or new physical entities are named in the abstract.

axioms (1)
  • domain assumption NAND flash cells require lower programming voltage for logical 1 than for logical 0
    Invoked to justify biasing codewords toward more 1s for energy and reliability gains.
invented entities (1)
  • Inverted Limited Weight Coding (ILWC) no independent evidence
    purpose: Encoding that adds flag bits to increase the fraction of 1s in stored data
    New coding construction introduced by the paper; no independent evidence supplied.

pith-pipeline@v0.9.0 · 5800 in / 1318 out tokens · 34474 ms · 2026-05-25T08:36:49.581390+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

41 extracted references · 41 canonical work pages

  1. [1]

    Advances in flash memory SSD technology for enterprise database applications,

    W. S. Lee, B. Moon, C. Park, “Advances in flash memory SSD technology for enterprise database applications,” Proceedings of the 2009 ACM SIGMOD International Conference on Management of data , pp. 863-870, 2009

    work page 2009

  2. [2]

    Solid state drive vs. hard disk drive price and performance study,

    V. Kasavajhala, “Solid state drive vs. hard disk drive price and performance study,” Proc. Dell Tech. White Paper, pp. 8-9, 2011

    work page 2011

  3. [3]

    Hybrid storage,

    R. Micheloni, L. Crippa, M. Picca, “Hybrid storage,” Inside Solid State Drives (SSDs), pp. 61-77, Springer Netherlands, 2013

    work page 2013

  4. [4]

    Introduction to flash memory,

    R. Bez, E. Camerlenghi, A. Modelli, A. Visconti, “Introduction to flash memory,” Proceedings of the IEEE. IEEE, pp. 489-502, 2003

    work page 2003

  5. [5]

    Inside solid state drives (SSDs),

    R. Micheloni, A. Marelli, K. Eshghi, “Inside solid state drives (SSDs),”. Springer Science & Business Media, 2012

    work page 2012

  6. [6]

    Novel co-design of NAND flash memory and NAND flash controller circuits for sub-30 nm low-power high-speed solid-state drives (SSD),

    K. Takeuchi, “Novel co-design of NAND flash memory and NAND flash controller circuits for sub-30 nm low-power high-speed solid-state drives (SSD),” IEEE Journal of solid-state circuits, pp. 124-125, 2009

    work page 2009

  7. [7]

    Systems, methods, and apparatus for controlling the power-on or boot sequence of an integrated circuit based on power harvesting conditions,

    B. Calhoun, D. Wentzloff, “Systems, methods, and apparatus for controlling the power-on or boot sequence of an integrated circuit based on power harvesting conditions,” U.S. Patent No. 9,558,008, 2017

    work page 2017

  8. [8]

    A Low -Power Hybrid Non -Volatile Cache with Asymmetric Coding,

    O. Hajihassani, A. Ahmadzadeh, M. Gavahi, M. R . Raei, D. Rahmati, S. Gorgin, “A Low -Power Hybrid Non -Volatile Cache with Asymmetric Coding,” 7th IEEE International Conference on Computer and Knowledge Engineering, pp. 277-282, 2017

    work page 2017

  9. [9]

    Power management in wireless tracking device operating with restricted power source,

    M. Turon, X. Yang, M. Dierks, “Power management in wireless tracking device operating with restricted power source, ” U.S. Patent No. 9,696,429. 4 Jul, 2017

    work page 2017

  10. [10]

    A case for flash memory SSD in enterprise database applications,

    S. W. Lee, B. Moon, C. Park, J. M. Kim, S. W. Kim “A case for flash memory SSD in enterprise database applications, ” Proceedings of the 2008 ACM SIGMOD international conference on Management of data , ACM, pp. 1075-1086, 2008

    work page 2008

  11. [11]

    Migrating server storage to SSDs: analysis of tradeoffs,

    D. Narayanan, E. Thereska, A. Donnelly, S. Elnikety, “Migrating server storage to SSDs: analysis of tradeoffs, ” Proceedings of the 4th ACM European conference on Computer systems, ACM, pp 145-158, 2009

    work page 2009

  12. [12]

    Understanding intrinsic characteristics and system implications of flash memory based solid state drives,

    F. Chen, D. A. Koufaty, X. Zhang, “Understanding intrinsic characteristics and system implications of flash memory based solid state drives,” ACM SIGMETRICS Performance Evaluation Review , 37(1) ACM, pp. 181-192, 2009

    work page 2009

  13. [13]

    NAND overview: from memory to systems,

    R. Micheloni, A. Marelli, A. Commodaro, “NAND overview: from memory to systems,” Inside NAND Flash Memories, Springer, pp. 19-53, 2010

    work page 2010

  14. [14]

    SSD reliability,

    C. Zambelli, P. Olivo, “SSD reliability,” Inside Solid State Drives (SSDs), pp. 203-231. Springer Netherlands, 2013

    work page 2013

  15. [15]

    Bus-invert coding for low-power I/O,

    M. R. Stan, W. P. Burleson, “Bus-invert coding for low-power I/O,” IEEE Transactions on very large scale integration VLSI systems, IEEE, pp. 49- 58, 1995

    work page 1995

  16. [16]

    Flashpower: A detailed power model for NAND flash memory,

    V. Mohan, S. Gurumurthi, R. M. Stan, “Flashpower: A detailed power model for NAND flash memory, ” Proceedings of the Confer ence on Design, Automation and Test in Europe , European Design and Automation Association, pp. 502-504, 2010

    work page 2010

  17. [17]

    Highly Reliable and Low Power SSD Using Asymmetric Coding and Stripe Bitline -Pattern Elimination Programming,

    S. Tanakamaru, C. Hung, K. Takeuchi, “Highly Reliable and Low Power SSD Using Asymmetric Coding and Stripe Bitline -Pattern Elimination Programming,” IEEE Journal of Solid-State Circuits 47(1), IEEE, pp. 85- 996, 2012

    work page 2012

  18. [18]

    A 4-Mbit NAND-EEPROM with tight programmed Vt distribution,

    T. Tanaka, M. Momodomi, Y. Iwata, Y. Tanaka, H. Oodaira, Y. Itoh, R. Shirota, K. Ohuchi, F. Masuoka, “A 4-Mbit NAND-EEPROM with tight programmed Vt distribution, ” Proceedings of Symposium on VLSI Circuits, pp. 105-106, 1990

    work page 1990

  19. [19]

    A 3.3 V 32 Mb NAND Flash memory with incremental step pulse programming scheme,

    K. D. Suh, B. H. Suh, Y. H. Lim, J. K. Kim, Y. J. Choi, Y. N. Koh, S. S. Lee, S. C. Kwon, B. S. Choi, J. S. Yum, J. H. Choi, “A 3.3 V 32 Mb NAND Flash memory with incremental step pulse programming scheme,” IEEE Journal of Solid -State Circuits, Vol. 30, No. 11, pp. 1149 -1156, 1995

    work page 1995

  20. [20]

    Exploring and exploiting the multilevel parallelism inside SSDs for improved performance and endurance,

    Y. Hu, H. Jiang, D. Feng, L. Tian, H. Luo, C. Ren, “Exploring and exploiting the multilevel parallelism inside SSDs for improved performance and endurance, ” IEEE Transactions on Computers (TC) , Vol. 62, No. 6, pp. 1141-1155, 2013

    work page 2013

  21. [21]

    Multi-channel architecture-based FTL for reliable and High performan ce SSD,

    J. W. Hsieh, H. Y. Lin, D. L. Yang, “Multi-channel architecture-based FTL for reliable and High performan ce SSD,” IEEE Transactions on Computers (TC), Vol. 63, No.12, pp. 3079-3091, 2014

    work page 2014

  22. [22]

    Less is more: trading a little bandwidth for ultra -low latency in the data center,

    M. Alizadeh, A. Kabbani, T. Edsall, B. Prabhakar, A. Vahdat, M. Yasuda, “Less is more: trading a little bandwidth for ultra -low latency in the data center,” Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation, USENIX Association, pp. 19-19, 2012. Type Technology size (nm) Dynamic power (w) Static power Delay (ns) Area...

    work page 2012

  23. [23]

    Bobtail: Avoiding Long Tails in the Cloud,

    Y. Xu, Z. Musgrave, B. Noble, M. Bailey, “Bobtail: Avoiding Long Tails in the Cloud,” NSDI, pp 329-342, 2013

    work page 2013

  24. [24]

    Providing QoS through Host Controlled Flash SSD,

    W. Shin, M. Kim, K. Kim, H. Y. Yeom, “Providing QoS through Host Controlled Flash SSD,” International Conference on Big Data and Smart Computing (BigComp), IEEE, pp. 111-117, 2015

    work page 2015

  25. [25]

    A semi - preemptive garbage collector for solid state drives,

    J. Lee, Y. Kim, G. M. Shipman, S. Oral, F. Wang, J. Kim, “A semi - preemptive garbage collector for solid state drives,” IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS) , IEEE, pp. 12-21, 2011

    work page 2011

  26. [26]

    Preemptible I/O scheduling of garbage collection for solid state drives,

    J. Lee, Y. Kim, G. M. Shipman, S. Oral, J. Kim, “Preemptible I/O scheduling of garbage collection for solid state drives,” IEEE Transactions on Computer -Aided Design of Integrated Circuits and Systems, 32(2). IEEE, pp. 247-260, 2013

    work page 2013

  27. [27]

    Efficient cache design for solid-state drives,

    M. Huang, O. Serres, V. K. Narayana, T. El -Ghazawi, G. Newby, “Efficient cache design for solid-state drives,” Proceedings of the 7th ACM international conference on Computing frontiers, ACM, pp. 41-50, 2010

    work page 2010

  28. [28]

    Page replacement algorithm based on counting bloom filter for NAND flash memory,

    J. Liu, S. Chen, G. Wang, T. Wu, “Page replacement algorithm based on counting bloom filter for NAND flash memory, ” IEEE Transactions on Consumer Electronics, 60(4), IEEE, pp. 636-643, 2014

    work page 2014

  29. [29]

    BPLRU: A Buffer Management Scheme for Improving Random Writes in Flash Storage,

    H. Kim, S. Ahn, “BPLRU: A Buffer Management Scheme for Improving Random Writes in Flash Storage,” FAST 8, pp. 1-14, 2008

    work page 2008

  30. [30]

    Built-in binary code inversion technique for on-chip flash memory sense amplifier with reduced read current consumption,

    D. Park, T. G. Kim, “Built-in binary code inversion technique for on-chip flash memory sense amplifier with reduced read current consumption, ” IEEE Transactions on Very Large Scale Integration (VLSI) Systems 22(5), IEEE, pp. 1187-119, 2014

    work page 2014

  31. [31]

    Understanding the impact of threshold voltage on MLC flash memo ry performance and reliability,

    W. Wang , T. Xie, D. Zhou, “ Understanding the impact of threshold voltage on MLC flash memo ry performance and reliability,” In Proceedings of the 28th ACM international conference on Supercomputing, ACM, pp. 201-210, 2014

    work page 2014

  32. [32]

    Stress induced leakage current limiting to scale down EEPROM tunnel oxide thickness,

    K. Naruke, S. Taguchi, M. Wada, “Stress induced leakage current limiting to scale down EEPROM tunnel oxide thickness, ” Electron Devices Meeting. IEDM'88. Technical Digest, IEEE, pp. 424-427, 1988

    work page 1988

  33. [33]

    Analytical percolation model for predicting anomalous charge loss in flash memories,

    R. Degraeve, F. Schuler, B. Kaczer, M. Lorenzini, D. Wellekens, P. Hendrickx, M. van Duuren, G. J. M. Dorm ans, J. Van Houdt, L. Haspeslagh, G. Groeseneken, “Analytical percolation model for predicting anomalous charge loss in flash memories,” IEEE Transactions on Electron Devices, 51(9). pp. 1392-1400, 2004

    work page 2004

  34. [34]

    Data retention in MLC NAND flash memory: Characterization, optimization, and recovery,

    Y. Cai, Y. Luo, E. F. Haratsch, K. Mai, O. Mutlu, “Data retention in MLC NAND flash memory: Characterization, optimization, and recovery, ” IEEE 21st International Symposium on High Performance Computer Architecture (HPCA), pp. 551-563, 2015

    work page 2015

  35. [35]

    A dual-mode NAND flash memory: 1-Gb multilevel and high -performance 512 -Mb single -level modes,

    T. Cho, Y.T. Lee, E.C. Kim, J.W. Lee, S. Choi, S. Lee, D.H. Kim, W.G. Han, Y.H. Lim, J.D. Lee, J.D. Choi, “A dual-mode NAND flash memory: 1-Gb multilevel and high -performance 512 -Mb single -level modes,” IEEE Journal of Solid-State Circuits 36, no. 11, pp. 1700-1706, 2001

    work page 2001

  36. [36]

    Bit err or rate in NAND flash memories,

    N. Mielke, T. Marquart, N. Wu, J. Kessenic h, H. Belgal, E. Schares, F. Trivedi, E. Goodness, L.R. Nevill, “ Bit err or rate in NAND flash memories,” In Reliability Physics Symposium , 2008. IRPS 2008. IEEE International, pp. 9-19. IEEE, 2008

    work page 2008

  37. [37]

    Characterizing flash memory: anomalies, observations, and applications,

    M. L. Grupp, A. M. Caulfield, J. Coburn, S. Swanson, E. Ya akobi, P. H. Siegel, J. K. Wolf, “Characterizing flash memory: anomalies, observations, and applications, ” 42nd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-42, pp. 24-33, 2009

    work page 2009

  38. [38]

    TID and SEE tests of an advanced 8 Gbit NAND -flash memory,

    H. Schmidt, D. Walter, F. Gliem , B. Nickson, R. Harboe -Sorensen, A. Virtanen, “TID and SEE tests of an advanced 8 Gbit NAND -flash memory,” Radiation Effects Data Workshop, IEEE. pp. 38-41, 2008

    work page 2008

  39. [39]

    SLC vs. MLC: An analysis of flash memory,

    K. Young, “SLC vs. MLC: An analysis of flash memory, ” Whitepaper, Super Talent Technology, Inc, 3, 2008

    work page 2008

  40. [40]

    Optimizing NAND flash-based SSDs via retention relaxation,

    R.S. Liu, C.L. Yang, W. Wu, “ Optimizing NAND flash-based SSDs via retention relaxation,” Target 11, no. 10, 2012

    work page 2012

  41. [41]

    A 117-mm/sup 2/3.3-V only 128-Mb multilevel NAND flash memory for mass storage applications,

    T.S. Jung, Y.J. Choi, K.D. Suh, B.H. Suh, J.K. Kim, Y.H. Lim, Y.N. Koh, J.W. Park, K.J. Lee, J.H. Park, K.T. Park, “ A 117-mm/sup 2/3.3-V only 128-Mb multilevel NAND flash memory for mass storage applications,” IEEE Journal of solid -state circuits 31, no. 11 , pp. 1575- 1583, 1996. Armin Ahmadzadeh received the B.Sc. and the M.Sc. degrees in computer eng...

    work page 1996