pith. sign in

arxiv: 2606.27319 · v1 · pith:X34KUU5Rnew · submitted 2026-06-25 · 💻 cs.ET · cs.AR

An Instruction Set Architecture for IMPLY-based Memristive Processing-in-Array

Liam Splittgerber , Fabian Seiler , Nima TaheriNejad This is my paper

Pith reviewed 2026-06-26 01:20 UTC · model grok-4.3

classification 💻 cs.ET cs.AR
keywords memristive computingIMPLY logicinstruction set architectureprocessing-in-memoryedge computingRISC-Vin-memory computingultra-low-power microcontrollers
0
0 comments X

The pith

Memristive crossbar arrays can host a complete RV32I-derived instruction set for standalone ultra-low-power microcontrollers.

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

The paper proposes an architecture that performs both data storage and computation inside the same memristive crossbar array so that no data needs to move between separate memory and processor. It adapts the RV32I instruction set for use with the IMPLY stateful logic operation and introduces a new addressing method that lets logic steps occur directly in the array. The design targets standalone general-purpose microcontrollers for edge devices rather than special-purpose accelerators, aiming to remove both the von Neumann bottleneck and the static leakage of volatile memory. If the mapping succeeds, the resulting hardware could run standard programs at far lower energy than conventional CMOS controllers. The work supplies a full implementation, circuit simulations that compare energy use, and a case study of an environmental sensor node to test practicality.

Core claim

The paper claims that a general-purpose instruction set architecture derived from RV32I can be realized entirely within a memristive crossbar array using the IMPLY stateful logic paradigm, supported by a novel addressing schema that enables computation in place, thereby providing a foundation for standalone ultra-low-power edge microcontrollers that unify storage and processing in non-volatile hardware.

What carries the argument

The IMPLY stateful logic paradigm, which executes logic operations directly on the resistance states of memristors, together with a custom addressing schema for the crossbar array.

If this is right

  • A complete implementation of the adapted RV32I instruction set is supplied for the memristive array.
  • Circuit-level simulations produce energy and other metrics that can be compared directly with traditional microcontrollers.
  • A case study demonstrates how the architecture could operate as a standalone controller in an intelligent environmental sensor node.
  • The novel addressing schema is required to keep all computation inside the crossbar without external data movement.

Where Pith is reading between the lines

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

  • Existing RISC-V software tools could be reused with only minor changes because the instruction set follows the RV32I standard.
  • The non-volatile storage might allow the device to resume instantly after power interruptions common in energy-harvesting edge nodes.
  • The same in-array approach could be tested with other stateful logic families to see whether they also support full general-purpose sets.

Load-bearing premise

That sequences of IMPLY operations can implement every instruction of a general-purpose set inside the array without external CMOS circuitry for control flow or temporary storage.

What would settle it

A concrete mapping or simulation test that shows at least one core RV32I instruction cannot be executed correctly using only IMPLY steps and the proposed addressing schema.

Figures

Figures reproduced from arXiv: 2606.27319 by Fabian Seiler, Liam Splittgerber, Nima TaheriNejad.

Figure 1
Figure 1. Figure 1: Circuit symbol of the memristor proposed in [4] [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Theoretical ideal model of a memristor [4, 5] [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The IMPLY logic gate using memristors Performing such a logical implication is done as shown in [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Memristive crossbar array for the serial topology [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: IMPLY operation across n rows the power used is solely for transferring data between computational and memory units [2, 3]. This fundamental bottleneck has generated a great deal of interest in computer architectures where the data never has to leave memory, and can be processed directly where the data resides. Using mem￾ristive crossbar arrays allows for just that. Data is stored in high-density memristiv… view at source ↗
Figure 6
Figure 6. Figure 6: RV32I instruction formats which further clarify the desired operation. The six instruction formats are as follows: 1. R-type (register-register): These instructions read data from two source registers (rs1 and rs2), perform a boolean or arithmetic operation, and store the result to a destination register (rd). 2. I-type (immediate): These instructions source one operand from the rs1 register, and the other… view at source ↗
Figure 7
Figure 7. Figure 7: Architecture of the crossbar array-based computer [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Control flow of the architecture Finally, the interface between the control logic and the crossbar array is handled by the row and column decoders and sense amplifiers. The decoders are implemented as demultiplexers, translating an address encoded as a binary value into a signal on the specific row and column lines corre￾15 [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Overview of the address bank and address format [PITH_FULL_IMAGE:figures/full_fig_p020_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Modified R-type encoding Since they all share an instruction format, the prerequisites for executing these instructions are uniform. The two source operands, denoted as A and B, occupy contiguous 32-bit blocks on the same row. Additionally, the execution of these algorithms requires one or more work memristors, which must also reside on the active row. In accordance with the architecture’s two-operand con… view at source ↗
Figure 11
Figure 11. Figure 11: Modified I-type encoding In the proposed architecture, the execution of these immediate instruc￾tions follows a two-stage process. First, the control logic reads the 12-bit immediate value from the instruction, sign-extends it to 32 bits, and writes this value directly into the crossbar array at the location designated for the first operand. This write operation is performed in two cycles: one cycle to as… view at source ↗
Figure 12
Figure 12. Figure 12: Modified U-type encoding 4.3. Load Store Instructions In conventional von Neumann architectures, load and store instructions are required for giving the CPU access to the data held within the memory. Since the CPU cannot operate directly on data residing in main memory, values must be explicitly moved into a limited set of registers. However, in the proposed in-memory architecture, storage and computation… view at source ↗
Figure 13
Figure 13. Figure 13: Modified J-type encoding 6.1.2. Instruction: jalr In RISC-V, the jalr instruction functions very similarly to the jal in￾struction, but instead of using a large immediate value as a PC-relative offset, it uses a register-based value. This enables the program to jump to an address calculated at runtime, which is often used for returning from a function or invoking a function pointer. The target address is … view at source ↗
Figure 14
Figure 14. Figure 14: Modified B-type encoding then evaluates the branch condition with a standard CMOS comparator. This is done primarily because performing 64 read operations (32 for each operand) and a CMOS comparison uses significantly less energy than all of the IMPLY operations involved in an in-memory comparison, making the proposed architecture better suited for edge computing. According to Kvatinsky et al. [18], readi… view at source ↗
Figure 15
Figure 15. Figure 15: Modified I-type encoding for address bank instructions [PITH_FULL_IMAGE:figures/full_fig_p040_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Modified U-type encoding for laui instruction The execution of the write operation follows the established two-phase write sequence, writing the ‘1’s and the ‘0’s in two separate steps. Similarly 41 [PITH_FULL_IMAGE:figures/full_fig_p041_16.png] view at source ↗
read the original abstract

The push towards expanded ultra-low-power edge computing necessitates hardware capable of operating under extremely strict energy constraints. Traditional Complementary Metal-Oxide-Semiconductor (CMOS) microcontrollers are fundamentally limited in this domain by the von Neumann bottleneck and by the static power leakage inherent to volatile memory. Memristive In-Memory Computing (IMC) offers a promising solution to these inefficiencies by unifying data storage and computation into a single non-volatile component. However, the State of the Art (SoA) predominantly focuses on accelerators designed to be a co-processor for data-intensive computation. This leaves the prospect of standalone, general-purpose IMC microcontroller architectures underexplored. This thesis proposes such an architecture tailored for ultra-low-power edge devices, alongside an instruction set closely derived from the RV32I standard. Using the IMPLY stateful logic paradigm, a complete implementation of the proposed instruction set is provided, and the novel addressing schema required to support computation in the memristive crossbar array is described as well. Then, the energy use and other circuit-level metrics of the proposed architecture are evaluated through simulation and compared against those of traditional microcontrollers. Finally, the functional viability of the design is demonstrated through an application case study, describing how the proposed design could be used in an intelligent environmental sensor node.

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 an instruction set architecture (ISA) for IMPLY-based memristive processing-in-array targeted at ultra-low-power edge devices. The ISA is closely derived from the RV32I standard. It claims a complete implementation using the IMPLY stateful logic paradigm, a novel addressing schema for computation in the memristive crossbar array, energy and circuit-level metric evaluations through simulation compared to traditional microcontrollers, and a functional demonstration via an application case study in an intelligent environmental sensor node.

Significance. If the proposed architecture and implementation are shown to be viable, this work would address a gap in the state of the art by exploring standalone general-purpose in-memory computing microcontrollers rather than co-processor accelerators. This could have implications for energy-constrained edge computing by mitigating the von Neumann bottleneck and static power leakage. The comparison to traditional CMOS microcontrollers and the case study would provide concrete evidence of potential benefits.

major comments (2)
  1. [Abstract] Abstract: The assertion of providing 'a complete implementation of the proposed instruction set' using the IMPLY paradigm is central to the claim of a functional standalone microcontroller, yet no mapping of RV32I instructions to IMPLY operations, no circuit diagrams, and no verification methods are provided, preventing assessment of whether the assumption that IMPLY can support a general-purpose ISA holds.
  2. [Abstract] Abstract: The energy evaluation through simulation and comparison to traditional microcontrollers is presented as a key contribution, but without details on the simulation setup, device models, or specific metrics (e.g., energy per instruction), it is impossible to determine if the results support the ultra-low-power claims.

Simulated Author's Rebuttal

2 responses · 1 unresolved

Thank you for the referee's comments. We address the major comments point by point.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The assertion of providing 'a complete implementation of the proposed instruction set' using the IMPLY paradigm is central to the claim of a functional standalone microcontroller, yet no mapping of RV32I instructions to IMPLY operations, no circuit diagrams, and no verification methods are provided, preventing assessment of whether the assumption that IMPLY can support a general-purpose ISA holds.

    Authors: The abstract is a concise summary of the manuscript's contributions and does not include the detailed technical elements mentioned. These are expected to be elaborated in the main sections of the paper. revision: no

  2. Referee: [Abstract] Abstract: The energy evaluation through simulation and comparison to traditional microcontrollers is presented as a key contribution, but without details on the simulation setup, device models, or specific metrics (e.g., energy per instruction), it is impossible to determine if the results support the ultra-low-power claims.

    Authors: Likewise, the abstract summarizes the energy evaluation without the specific setup details, which would be provided in the evaluation section of the full manuscript. revision: no

standing simulated objections not resolved
  • Full manuscript text is not available beyond the abstract, so it is not possible to verify if the detailed mappings, circuit diagrams, verification methods, simulation setups, device models, and metrics are present in the paper.

Circularity Check

0 steps flagged

No circularity detectable; abstract-only text provides no derivations

full rationale

The supplied document contains only the abstract, which outlines a high-level proposal for an IMPLY-based ISA derived from RV32I along with a novel addressing schema and simulation-based evaluation. No equations, circuit mappings, operation sequences, or derivation chains are present to inspect. Consequently, none of the enumerated circularity patterns (self-definitional, fitted-input prediction, self-citation load-bearing, etc.) can be exhibited by quoting specific text and showing a reduction to inputs. The central claim of functional viability remains an unelaborated assertion rather than a chain that collapses by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no information on free parameters, axioms, or invented entities; insufficient detail for ledger population.

pith-pipeline@v0.9.1-grok · 5739 in / 1003 out tokens · 38203 ms · 2026-06-26T01:20:40.932057+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

23 extracted references · 11 canonical work pages

  1. [1]

    U. G. Shinde, T. Luhmann, J. Klueppel, T. Steinmann, L. M. Comella, W. Yu, S. J. Rupitsch, P. Woias, Task-specific energy profiling for microcontroller selection in energy-autonomous wireless sensor nodes, IEEE Sensors Journal 25 (21) (2025) 40609–40622. doi:10.1109/JSEN.2025.3609827

    work page doi:10.1109/jsen.2025.3609827 2025

  2. [2]

    Computing’s energy problem (and what we can do about it)

    M. Horowitz, 1.1 computing’s energy problem (and what we can do about it), in: 2014 IEEE International Solid-State Circuits Conference Digest of Technical Papers (ISSCC), 2014, pp. 10–14. doi:10.1109/ISSCC.2014.6757323

    work page doi:10.1109/isscc.2014.6757323 2014

  3. [3]

    TaheriNejad, In-memory computing: Global energy consumption, carbon footprint, technology, and products status quo (2024) 1–6

    N. TaheriNejad, In-memory computing: Global energy consumption, carbon footprint, technology, and products status quo (2024) 1–6

    2024

  4. [4]

    Chua, Memristor-the missing circuit element, IEEE Transactions on Circuit Theory 18 (5) (1971) 507–519

    L. Chua, Memristor-the missing circuit element, IEEE Transactions on Circuit Theory 18 (5) (1971) 507–519. doi:10.1109/TCT.1971.1083337

    work page doi:10.1109/tct.1971.1083337 1971

  5. [5]

    Borghetti, G

    J. Borghetti, G. Snider, P. Kuekes, J. J. Yang, D. Stewart, S. Williams, Memristive switches enable stateful logic operations via material impli- cation, Nature 464 (2010) 873–6. doi:10.1038/nature08940

    work page doi:10.1038/nature08940 2010

  6. [6]

    M. S. Q. Truong, E. Chen, D. Su, L. Shen, A. Glass, L. R. Carley, J. A. Bain, S. Ghose, Racer: Bit-pipelined processing using resistive memory, in: MICRO-54: 54th Annual IEEE/ACM International Symposium on Microarchitecture, MICRO ’21, Association for Computing Machinery, New York, NY, USA, 2021, p. 100–116. doi:10.1145/3466752.3480071. URLhttps://doi.org...

    work page doi:10.1145/3466752.3480071 2021

  7. [7]

    Leitersdorf, D

    O. Leitersdorf, D. Leitersdorf, J. Gal, M. M. Dahan, R. Ronen, S. Kvatinsky, Aritpim: High-throughput in-memory arithmetic, IEEE Transactions on Emerging Topics in Computing 11 (2022) 720–735. URLhttps://api.semanticscholar.org/CorpusID:249538709 55

    2022

  8. [8]

    Seiler, P

    F. Seiler, P. M. Hinkel, A. Jantsch, N. TaheriNejad, Atomic: Automatic tool for memristive imply based circuit-level simulation and validation, in: 2025 21st International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD), 2025, pp. 1–4

    2025

  9. [9]

    Kvatinsky, A

    S. Kvatinsky, A. Kolodny, U. C. Weiser, E. G. Friedman, Memristor- based imply logic design procedure, in: 2011 IEEE 29th Interna- tional Conference on Computer Design (ICCD), 2011, pp. 142–147. doi:10.1109/ICCD.2011.6081389

    work page doi:10.1109/iccd.2011.6081389 2011

  10. [10]

    D. B. Strukov, G. Snider, D. R. Stewart, R. S. Williams, The missing memristor found, Nature 453 (2008) 80–83

    2008

  11. [11]

    Kvatinsky, D

    S. Kvatinsky, D. Belousov, S. Liman, G. Satat, N. Wald, E. G. Fried- man, A. Kolodny, U. C. Weiser, Magic—memristor-aided logic, IEEE Transactions on Circuits and Systems II: Express Briefs 61 (11) (2014) 895–899. doi:10.1109/TCSII.2014.2357292

    work page doi:10.1109/tcsii.2014.2357292 2014

  12. [12]

    Gupta, M

    S. Gupta, M. Imani, T. Rosing, Felix: Fast and energy-efficient logic in memory, in: 2018 IEEE/ACM International Conference on Computer- Aided Design (ICCAD), 2018, pp. 1–7. doi:10.1145/3240765.3240811

    work page doi:10.1145/3240765.3240811 2018

  13. [13]

    Radakovits, N

    D. Radakovits, N. Taherinejad, Behavioral leakage and inter-cycle vari- ability emulator model for rerams (BELIEVER), CoRR abs/2103.04179 (2021). arXiv:2103.04179. URLhttps://arxiv.org/abs/2103.04179

    arXiv 2021

  14. [14]

    Lehtonen, M

    E. Lehtonen, M. Laiho, Stateful implication logic with memristors, 2009 IEEE/ACM International Symposium on Nanoscale Architectures (2009) 33–36

    2009

  15. [15]

    Waterman, Design of the risc-v instruction set architecture, Ph.D

    A. Waterman, Design of the risc-v instruction set architecture, Ph.D. thesis, EECSDepartment, UniversityofCalifornia, Berkeley(Jan2016). URLhttp://www2.eecs.berkeley.edu/Pubs/TechRpts/2016/EECS-2016-1.html

    2016

  16. [16]

    Seiler, N

    F. Seiler, N. TaheriNejad, An improved serial imply adder algorithm for efficient neural network applications, in: 2025 IEEE 16th Latin America Symposium on Circuits and Systems (LASCAS), 2025, pp. 1–5

    2025

  17. [17]

    Kaushik, B

    N. Kaushik, B. Srinivasu, Implementation of imply-based memristive subtractor, in: 2022 IEEE 22nd International Conference on Nanotechnology (NANO), 2022, pp. 523–526. doi:10.1109/NANO54668.2022.9928713. 56

    work page doi:10.1109/nano54668.2022.9928713 2022

  18. [18]

    Kvatinsky, G

    S. Kvatinsky, G. Satat, N. Wald, E. G. Friedman, A. Kolodny, U. C. Weiser, Memristor-based material implication (imply) logic: Design principles and methodologies, IEEE Transactions on Very Large Scale Integration (VLSI) Systems 22 (10) (2014) 2054–2066. doi:10.1109/TVLSI.2013.2282132

    work page doi:10.1109/tvlsi.2013.2282132 2014

  19. [19]

    Brum, Pyltspice (2024)

    N. Brum, Pyltspice (2024). URLhttps://github.com/nunobrum/PyLTSpice

    2024

  20. [20]

    Radakovits, M

    D. Radakovits, M. Jungwirth, S. M. Laube, N. TaheriNe- jad, Second (v2.0) LTSpice implementation of VTEAM, https://www.ict.tuwien.ac.at/staff/taherinejad/projects/ memristor/files/vteam2.asc, https://www.ict.tuwien.ac.at/staff/ taherinejad/projects/memristor/files/vteam2.asy (September 2019)

    2019

  21. [21]

    Kvatinsky, M

    S. Kvatinsky, M. Ramadan, E. G. Friedman, A. Kolodny, Vteam: A general model for voltage-controlled memristors, IEEE Transactions on Circuits and Systems II: Express Briefs 62 (8) (2015) 786–790. doi:10.1109/TCSII.2015.2433536

    work page doi:10.1109/tcsii.2015.2433536 2015

  22. [22]

    URLhttps://knowm.org/downloads/Knowm_Memristors.pdf

    Knowm sdc memristors, last accessed Feb 2024 (knowm.org/downloads/Knowm_Memristors.pdf). URLhttps://knowm.org/downloads/Knowm_Memristors.pdf

    2024

  23. [23]

    Column A

    Texas Instruments, TMP102 Low-Power Digital Temperature Sensor With SMBus and Two-Wire Serial Interface in SOT563, datasheet SBOS451J (Revised Dec. 2018) (Dec 2018). URLhttps://www.ti.com/lit/ds/symlink/tmp102.pdf Appendix A. Assembly for Temperature Sensor Node 1#|================================================================| 2#|======================...

    2018