pith. sign in

arxiv: 2509.08193 · v2 · submitted 2025-09-09 · 💻 cs.AR · cs.AI· cs.ET

Lifetime-Aware Design for Item-Level Intelligence at the Extreme Edge

Shvetank Prakash , Andrew Cheng , Olof Kindgren , Ashiq Ahamed , Graham Knight , Jed Kufel , Francisco Rodriguez , Arya Tschand
show 8 more authors
David Kong Mariam Elgamal Jerry Huang Emma Chen Gage Hills Richard Price Emre Ozer Vijay Janapa Reddi
This is my paper

Pith reviewed 2026-05-18 17:33 UTC · model grok-4.3

classification 💻 cs.AR cs.AIcs.ET
keywords item-level intelligenceflexible electronicscarbon footprintlifetime-aware designRISC-Vextreme edgesustainabilitylow-precision computing
0
0 comments X

The pith

Lifetime-aware design for flexible electronics in disposable items cuts carbon footprint by 1.62X through architecture and 14.5X through algorithms at trillion-item scales.

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

The paper introduces FlexiFlow, a framework for embedding simple computation directly into single-use products such as food packaging or medical patches using low-cost flexible electronics instead of silicon. Its central argument is that these item-level intelligence applications show operational lifetimes that vary by a factor of 1000, so the relative weight of manufacturing emissions versus runtime emissions shifts enough to change which processor designs are optimal. The work supplies benchmark workloads, bit-width-optimized RISC-V cores, and a carbon model that picks architectures according to each application's expected lifetime. If correct, the approach shows how to keep total emissions low even when trillions of such devices are deployed.

Core claim

FlexiFlow models the trade-off between embodied carbon from manufacturing and operational carbon from use, parameterized by each application's specific lifetime. It supplies FlexiBench workloads drawn from sustainability tasks, FlexiBits RISC-V cores using 1-, 4-, or 8-bit datapaths that improve energy efficiency per execution by 2.65X to 3.50X, and a selection model that chooses the lowest-carbon architecture for given deployment lifetimes. The framework demonstrates that lifetime-aware microarchitectural choices reduce total carbon by 1.62X while algorithmic decisions reduce it by 14.5X, and it validates the flow with an open-source tape-out on a flexible-electronics PDK that reaches 30.9k

What carries the argument

A carbon-aware model that selects optimal architectures by balancing embodied versus operational emissions according to application-specific operational lifetimes.

If this is right

  • Shorter-lifetime applications favor simpler low-bit-width cores that minimize embodied carbon.
  • Algorithm-level choices produce larger carbon reductions than microarchitectural changes alone.
  • Flexible electronics become practical for extreme-edge intelligence once designs are tuned to limited lifetimes.
  • Trillion-item volumes make even modest per-device lifetime-specific optimizations globally material.
  • Open-source tape-out at 30.9 kHz confirms that the selected low-gate-count designs are manufacturable with current flexible processes.

Where Pith is reading between the lines

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

  • The same lifetime-parameterized carbon model could be applied to other high-volume, low-cost substrates beyond flexible electronics.
  • Designers of future disposable sensors might need to publish expected lifetime ranges so that downstream architects can select appropriate bit widths.
  • At trillion-item scale, recycling or end-of-life carbon accounting could interact with the embodied-versus-operational split modeled here.
  • The framework suggests that conventional uniform-lifetime assumptions in edge-computing literature may systematically overestimate optimal complexity for short-lived uses.

Load-bearing premise

The claim that item-level intelligence applications exhibit up to 1000X variation in operational lifetime, which is treated as the factor that forces different architecture choices at large scale.

What would settle it

Direct measurement of operational lifetimes for a representative set of real ILI applications, such as spoilage sensors and health patches, followed by recalculation of the carbon trade-off curves to check whether the reported 1.62X and 14.5X reductions still appear.

Figures

Figures reproduced from arXiv: 2509.08193 by Andrew Cheng, Arya Tschand, Ashiq Ahamed, David Kong, Emma Chen, Emre Ozer, Francisco Rodriguez, Gage Hills, Graham Knight, Jed Kufel, Jerry Huang, Mariam Elgamal, Olof Kindgren, Richard Price, Shvetank Prakash, Vijay Janapa Reddi.

Figure 1
Figure 1. Figure 1: Item-level Intelligence (ILI) in comparison with traditional mobile/edge computing and cloud computing. ILI targets unprecedented scale. The volume involves tril￾lions of units annually compared to millions for mobile de￾vices [43, 44, 63, 64, 89–92]. Even at scale, silicon-based mi￾crocontrollers cost tens of cents to dollars per unit [36, 42], which is prohibitive for integration into low-margin con￾sume… view at source ↗
Figure 2
Figure 2. Figure 2: Computational patterns of FlexiBench workloads. better supported today. Furthermore, SRAM consumes sig￾nificantly higher power consumption than LPROM, making VM size a key factor influencing lifetime-aware optimization. FlexiBench’s diversity also helps identify the techno￾logical gaps to be addressed to enable systematic progress towards ILI. We observe that the memory requirements of some benchmarks exce… view at source ↗
Figure 3
Figure 3. Figure 3: illustrates our template microarchitecture that scales across these datapath widths. The key architectural contribution is the clean separation between width-independent Control Unit Control Status Registers Arithmetic Logic Unit Buffer Register #2 Memory Interface Buffer Register #1 Immediate Decoder Data Bus/ Extension Support Decoder State Machine Instruction Bus Address Register File Interface Data Bus… view at source ↗
Figure 4
Figure 4. Figure 4: FlexiFlow takes inputs spanning the computing stack including user specifications, architecture design parameters, and foundry data to identify the carbon-optimal design using profiling data in conjunction with its lifetime-aware model. 4.4 Performance, Power, and Area Comparison We perform standard PPA analysis of our FlexiBits cores on FlexiBench, leaving extended details to Appendix B.1. Performance. Wo… view at source ↗
Figure 5
Figure 5. Figure 5: Carbon-optimal system selection for ILI applications generated using FlexiFlow depending on lifetime and program execution frequency. Red stars correspond to the example use cases from [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Pareto frontier of classification accuracy vs. total carbon footprint for a 1-year lifetime across different soft￾ware implementations of food spoilage detection. first-class concern? We examine food spoilage detection, where prior work has proposed multiple algorithms: Deci￾sion Trees (DT) [48], k-Nearest Neighbors (KNN) [30, 48], Logistic Regression (LR) [30], and MLPs [30]. We evaluate multiple configur… view at source ↗
Figure 7
Figure 7. Figure 7: Open-source tape out from FlexiFlow. to estimate how often integrating ILI will actually save a slab of beef that would have been wasted4 due to uncontrollable factors (e.g., human behavior), this table sweeps various effectiveness-rates of ILI (e.g. 100% means all to-be-wasted meat is saved). To ground the at-scale savings numbers, we additionally calculate the equivalent number of yearly car emissions th… view at source ↗
Figure 8
Figure 8. Figure 8: Cycle-level execution time of FlexiBench workloads across the FlexiBits microprocessors. *The tree tracking (TT) workload was simulated using instruction-level simulation due to its extremely high runtime. Workload SERV QERV HERV Food Spoilage Detection ✓ ✓ ✓ Cardiotocography ✓ ✓ ✓ Arrhythmia Detection ✗ ✗ ✗ Water Quality Monitoring ✓ ✓ ✓ HVAC Control ✓ ✓ ✓ Package Tracking ✓ ✓ ✓ Gesture Recognition ✗ ✗ ✗ … view at source ↗
Figure 9
Figure 9. Figure 9: Scaling of power, area, runtime, and energy effi￾ciency of FlexiBits cores. Runtime scaling is the geometric mean of scaling across all FlexiBench workloads. 3.50× for QERV and HERV, respectively). This energy-efficiency benefit quantitatively supports the qualitative intuition de￾scribed in Section 5.5: Increasing the datapath width will incur higher area and embodied carbon footprint, but it will also lo… view at source ↗
Figure 10
Figure 10. Figure 10: Normalized breakdown of power for each work￾load between compute (FlexiBits) and memory. As LPROM consumes negligible power, memory power is only from SRAM requirements. FlexiBits Datapath Area (mm2 ) Area Over￾head Power (mW) Power Over￾head SERV (1-bit) 2.93 1× 17.75 1× QERV (4-bit) 3.68 1.26× 21.07 1.19× HERV (8-bit) 4.50 1.54× 24.99 1.41× [PITH_FULL_IMAGE:figures/full_fig_p020_10.png] view at source ↗
Figure 12
Figure 12. Figure 12 [PITH_FULL_IMAGE:figures/full_fig_p020_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Energy ablation study on the Air Pollution Moni￾toring workload. Plots ordered from highest carbon intensity (coal) on the left to lowest carbon intensity (solar) on the right. As shown, higher carbon intensities favor the larger, more energy efficient QERV and HERV. B.3.1 Instruction Mix. The instruction mix of each work￾load typically will not significantly impact the carbon ac￾counting of FlexiFlow. Th… view at source ↗
Figure 14
Figure 14. Figure 14: Open-source physical implementation and wafer testing of the FlexiBits-based SoC. To this end, we integrate Pragmatic Semiconductor’s pro￾cess development kit (PDK) with OpenROAD [3], an open￾source EDA toolchain. To support this integration, custom standard cells were developed to enable correct placement of sequential elements such as D flip-flops. Additional modifi￾cations were made to the technology l… view at source ↗
read the original abstract

We present FlexiFlow, a lifetime-aware design framework for item-level intelligence (ILI) where computation is integrated directly into disposable products like food packaging and medical patches. Our framework leverages natively flexible electronics which offer significantly lower costs than silicon but are limited to kHz speeds and several thousands of gates. Our insight is that unlike traditional computing with more uniform deployment patterns, ILI applications exhibit 1000X variation in operational lifetime, fundamentally changing optimal architectural design decisions when considering trillion-item deployment scales. To enable holistic design and optimization, we model the trade-offs between embodied carbon footprint and operational carbon footprint based on application-specific lifetimes. The framework includes: (1) FlexiBench, a workload suite targeting sustainability applications from spoilage detection to health monitoring; (2) FlexiBits, area-optimized RISC-V cores with 1/4/8-bit datapaths achieving 2.65X to 3.50X better energy efficiency per workload execution; and (3) a carbon-aware model that selects optimal architectures based on deployment characteristics. We show that lifetime-aware microarchitectural design can reduce carbon footprint by 1.62X, while algorithmic decisions can reduce carbon footprint by 14.5X. We validate our approach through the first tape-out using a PDK for flexible electronics with fully open-source tools, achieving 30.9kHz operation. FlexiFlow enables exploration of computing at the Extreme Edge where conventional design methodologies must be reevaluated to account for new constraints and considerations.

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 presents FlexiFlow, a lifetime-aware design framework for item-level intelligence (ILI) at the extreme edge using natively flexible electronics. It introduces FlexiBench (workloads from spoilage detection to health monitoring), FlexiBits (area-optimized RISC-V cores with 1/4/8-bit datapaths claiming 2.65X–3.50X better energy efficiency), and a carbon-aware model that selects architectures based on application-specific lifetimes asserted to vary by 1000X. The central claims are that lifetime-aware microarchitectural design reduces carbon footprint by 1.62X and algorithmic decisions by 14.5X at trillion-item scales, validated by a tape-out in a flexible-electronics PDK using fully open-source tools that achieves 30.9 kHz operation.

Significance. If the lifetime-variation premise and carbon model are substantiated, the work would be significant for computer architecture targeting sustainable, disposable trillion-scale applications, where embodied vs. operational carbon trade-offs differ markedly from conventional computing. The concrete tape-out results and open-source tool flow constitute a clear strength, providing a reproducible demonstration of feasibility for flexible-electronics microarchitecture. The FlexiBench and FlexiBits artifacts could serve as useful baselines for future extreme-edge sustainability studies.

major comments (2)
  1. [Introduction and carbon-aware model description] The 1000X operational-lifetime variation is load-bearing for the optimality claims yet is presented as an input without traceable data sources, sensitivity analysis, or per-item statistical modeling. If the actual range is closer to 10–100X, the differential dominance of operational carbon across architectures would shrink and the reported 1.62X microarchitectural saving would not necessarily follow.
  2. [Carbon-aware model] No equation or table shows how lifetime enters the embodied-vs-operational carbon calculation or how the 1.62X and 14.5X factors were derived under the stated lifetime distribution. Without these, it is impossible to confirm that the reductions are independent predictions rather than partly defined by the same lifetime values used to select the architectures.
minor comments (2)
  1. [Abstract] The abstract states concrete speedups and carbon reductions without error bars or confidence intervals; adding these would improve clarity of the quantitative claims.
  2. [Tape-out validation] The tape-out section would benefit from a brief comparison table against prior flexible-electronics implementations to contextualize the 30.9 kHz result.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the valuable feedback on our manuscript. We have carefully considered the comments regarding the lifetime variation and the carbon model and provide point-by-point responses below. Revisions have been made to address the concerns.

read point-by-point responses

  1. Referee: The 1000X operational-lifetime variation is load-bearing for the optimality claims yet is presented as an input without traceable data sources, sensitivity analysis, or per-item statistical modeling. If the actual range is closer to 10–100X, the differential dominance of operational carbon across architectures would shrink and the reported 1.62X microarchitectural saving would not necessarily follow.

    Authors: We agree that the 1000X variation requires better substantiation. The revised manuscript now includes references to studies on ILI lifetimes for different applications, such as short-lived food sensors and longer-term medical devices, supporting the wide range. We have incorporated a sensitivity analysis showing that the reported savings are robust down to 100X variation. Regarding per-item statistical modeling, our framework uses representative lifetime distributions per application class rather than individual item modeling, which we have clarified. revision: yes

  2. Referee: No equation or table shows how lifetime enters the embodied-vs-operational carbon calculation or how the 1.62X and 14.5X factors were derived under the stated lifetime distribution. Without these, it is impossible to confirm that the reductions are independent predictions rather than partly defined by the same lifetime values used to select the architectures.

    Authors: We acknowledge the need for explicit derivation. In the updated manuscript, we have added the carbon calculation equation: total carbon footprint = embodied carbon + (operational energy per execution × execution rate × lifetime). A new table details the breakdown for each FlexiBits configuration under the lifetime distribution, deriving the 1.62X microarchitectural and 14.5X algorithmic reductions by comparing against non-lifetime-aware baselines. These are computed post-selection to validate the benefits. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper introduces FlexiFlow as a modeling framework that takes application-specific lifetimes (asserted as exhibiting 1000X variation) as an explicit input parameter to compute embodied-vs-operational carbon trade-offs and select among FlexiBits architectures. The reported 1.62X microarchitectural and 14.5X algorithmic carbon reductions are presented as computed outcomes of applying this model to the FlexiBench workloads and trillion-item scale assumptions, not as quantities that are defined by or statistically forced back into the lifetime inputs. No self-citations, uniqueness theorems, or ansatzes are described as load-bearing; the 1000X premise is an external modeling choice rather than a result derived from the framework itself. The chain from workload definition through core energy measurements to carbon-aware selection remains independently falsifiable via the open-source tape-out and per-workload execution data.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on standard assumptions about flexible electronics limits and carbon accounting; the 1000X lifetime variation is presented as an empirical observation rather than a fitted parameter.

free parameters (1)
  • lifetime variation range
    1000X spread across ILI applications is used as the central driver for architecture selection.
axioms (1)
  • domain assumption Natively flexible electronics are limited to kHz speeds and several thousands of gates
    Invoked to justify why conventional silicon design flows must be reevaluated.

pith-pipeline@v0.9.0 · 5862 in / 1330 out tokens · 50485 ms · 2026-05-18T17:33:46.941358+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

125 extracted references · 125 canonical work pages

  1. [1]

    Kofi Sarpong Adu-Manu, Cristiano Tapparello, Wendi Heinzelman, Ferdinand Apietu Katsriku, and Jamal-Deen Abdulai. 2017. Water Quality Monitoring Using Wireless Sensor Networks: Current Trends and Future Research Directions.ACM Transactions on Sensor Net- works13, 1 (Feb. 2017), 1–41. doi:10.1145/3005719

    work page doi:10.1145/3005719 2017

  2. [2]

    Ahamed, C

    A. Ahamed, C. Anand, and J. Young. 2023. Life Cycle Analysis of FlexICs. White Paper, Pragmatic. [Online]. Available:https://www. pragmaticsemi.com

    work page 2023

  3. [3]

    Tutu Ajayi, Vidya A Chhabria, Mateus Fogaça, Soheil Hashemi, Ab- delrahman Hosny, Andrew B Kahng, Minsoo Kim, Jeongsup Lee, Uday Mallappa, Marina Neseem, et al. 2019. Toward an open-source digital flow: First learnings from the openroad project. InProceedings of the 56th Annual Design Automation Conference 2019. 1–4

    work page 2019

  4. [4]

    Tahoori, and Jörg Henkel

    Giorgos Armeniakos, Georgios Zervakis, Dimitrios Soudris, Mehdi B. Tahoori, and Jörg Henkel. 2022. Cross-Layer Approximation For Printed Machine Learning Circuits. In2022 Design, Automation & Test in Europe Conference & Exhibition (DATE). 190–195. doi:10.23919/ DATE54114.2022.9774689

    work page arXiv 2022

  5. [5]

    Paddy M Barrett, Ravi Komatireddy, Sharon Haaser, Sarah Topol, Judith Sheard, Jackie Encinas, Angela J Fought, and Eric J Topol. 2014. Comparison of 24-hour Holter monitoring with 14-day novel adhe- sive patch electrocardiographic monitoring.The American journal of medicine127, 1 (2014), 95–e11

    work page 2014

  6. [6]

    A Belmonte, H Oh, N Rassoul, GL Donadio, J Mitard, H Dekkers, R Delhougne, S Subhechha, A Chasin, MJ Van Setten, et al . 2020. Capacitor-less, long-retention (> 400s) DRAM cell paving the way towards low-power and high-density monolithic 3D DRAM. In2020 IEEE International Electron Devices Meeting (IEDM). IEEE, 28–2

    work page 2020

  7. [7]

    John Biggs, James Myers, Jedrzej Kufel, Emre Ozer, Simon Craske, Antony Sou, Catherine Ramsdale, Ken Williamson, Richard Price, and Scott White. 2021. A natively flexible 32-bit Arm microprocessor. Nature595, 7868 (2021), 532–536

    work page 2021

  8. [8]

    Nathaniel Bleier, Calvin Lee, Francisco Rodriguez, Antony Sou, Scott White, and Rakesh Kumar. 2022. FlexiCores: low footprint, high yield, field reprogrammable flexible microprocessors. InProceedings of the 49th Annual International Symposium on Computer Architecture. 831–846

    work page 2022

  9. [9]

    Nathaniel Bleier, M Husnain Mubarik, Suman Balaji, Francisco Ro- driguez, Antony Sou, Scott White, and Rakesh Kumar. 2023. Ex- ploiting short application lifetimes for low cost hardware encryption in flexible electronics. In2023 Design, Automation & Test in Europe Conference & Exhibition (DATE). IEEE, 1–6

    work page 2023

  10. [10]

    Nathaniel Bleier, Abigail Wezelis, Lav Varshney, and Rakesh Kumar

    work page

  11. [11]

    Rowpress: Amplifying read disturbance in modern dram chips,

    Programmable Olfactory Computing. InProceedings of the 50th Annual International Symposium on Computer Architecture (ISCA ’23). Association for Computing Machinery, New York, NY, USA, 1–14. doi:10.1145/3579371.3589061

    work page doi:10.1145/3579371.3589061

  12. [12]

    Buzby, Hodan Farah-Wells, and Jeffrey Hyman

    Jean C. Buzby, Hodan Farah-Wells, and Jeffrey Hyman. 2014. The Estimated Amount, Value, and Calories of Postharvest Food Losses at the Retail and Consumer Levels in the United States.USDA-ERS Economic Information Bulletin121 (2014)

    work page 2014

  13. [13]

    Cadence. 2025. Genus Synthesis Solution. https://www.cadence.com/en_US/home/tools/digital-design- and-signoff/synthesis/genus-synthesis-solution.html

    work page 2025

  14. [14]

    Campos and J

    D. Campos and J. Bernardes. 2000. Cardiotocography. UCI Machine Learning Repository. DOI: https://doi.org/10.24432/C51S4N

    work page doi:10.24432/c51s4n 2000

  15. [15]

    Candanedo and Véronique Feldheim

    Luis M. Candanedo and Véronique Feldheim. 2016. Accurate Occu- pancy Detection of an Office Room from Light, Temperature, Humid- ity and CO2 Measurements Using Statistical Learning Models.Energy and Buildings112 (Jan. 2016), 28–39. doi:10.1016/j.enbuild.2015.11.071

    work page doi:10.1016/j.enbuild.2015.11.071 2016

  16. [16]

    Hikmet Çeliker, Wim Dehaene, and Kris Myny. 2024. Multi-Project Wafers for Flexible Thin-Film Electronics by Independent Foundries. Nature629, 8011 (May 2024), 335–340. doi:10.1038/s41586-024-07306- 2

    work page doi:10.1038/s41586-024-07306- 2024

  17. [17]

    Hikmet Çeliker, Antony Sou, Brian Cobb, Wim Dehaene, and Kris Myny. 2022. Flex6502: a flexible 8b microprocessor in 0.8 𝜇m metal- oxide thin-film transistor technology implemented with a complete digital design flow running complex assembly code. In2022 IEEE International Solid-State Circuits Conference (ISSCC), Vol. 65. IEEE, 272–274

    work page 2022

  18. [18]

    Emma Chen, Shvetank Prakash, Vijay Janapa Reddi, David Kim, and Pranav Rajpurkar. 2025. A framework for integrating artificial intelligence for clinical care with continuous therapeutic monitoring. Nature Biomedical Engineering9, 4 (2025), 445–454

    work page 2025

  19. [19]

    Guorui Chen, Yongzhong Li, Michael Bick, and Jun Chen. 2020. Smart textiles for electricity generation.Chemical Reviews120, 8 (2020), 3668–3720

    work page 2020

  20. [20]

    Kunigunde Cherenack and Liesbeth Van Pieterson. 2012. Smart textiles: Challenges and opportunities.Journal of Applied Physics 112, 9 (2012)

    work page 2012

  21. [21]

    Coder’s Cafe. 2022. CodersCafeTech/Package-Tracker. Coder’s Cafe

    work page 2022

  22. [22]

    De Vito, E

    S. De Vito, E. Massera, M. Piga, L. Martinotto, and G. Di Francia. 2008. On Field Calibration of an Electronic Nose for Benzene Estimation in an Urban Pollution Monitoring Scenario.Sensors and Actuators B: Chemical129, 2 (Feb. 2008), 750–757. doi:10.1016/j.snb.2007.09.060

    work page doi:10.1016/j.snb.2007.09.060 2008

  23. [23]

    Priyanka deSouza, Karoline Barkjohn, Andrea Clements, Jenny Lee, Ralph Kahn, Ben Crawford, and Patrick Kinney. 2023. An analysis of degradation in low-cost particulate matter sensors.Environmental science: atmospheres3, 3 (2023), 521–536

    work page 2023

  24. [24]

    Asbjørn Djupdal, Magnus Själander, Magnus Jahre, Snorre Aunet, and Trond Ytterdal. 2025. Optimizing Energy Efficiency in Subthreshold RISC-V Cores. arXiv:2502.06588 [cs] doi:10.48550/arXiv.2502.06588

    work page doi:10.48550/arxiv.2502.06588 2025

  25. [25]

    Lieven Eeckhout. 2024. FOCAL: A First-Order Carbon Model to Assess Processor Sustainability. InProceedings of the 29th ACM Inter- national Conference on Architectural Support for Programming Lan- guages and Operating Systems, Volume 2 (ASPLOS ’24, Vol. 2). As- sociation for Computing Machinery, New York, NY, USA, 401–415. doi:10.1145/3620665.3640415

    work page doi:10.1145/3620665.3640415 2024

  26. [26]

    Lieven Eeckhout. 2024. Rip geomean speedup use equal-work (or equal-time) harmonic mean speedup instead.IEEE Computer Archi- tecture Letters23, 1 (2024), 78–82

    work page 2024

  27. [27]

    eeNews. 2025. How much ‘Moore’ do we need? New technologies for the extreme edge! [Online]. Available: https://www.eenewseurope.com/en/how-much-moore-do-we- 12 need-new-technologies-for-the-extreme-edge/

    work page 2025

  28. [28]

    Mariam Elgamal, Doug Carmean, Elnaz Ansari, Okay Zed, Ramesh Peri, Srilatha Manne, Udit Gupta, Gu-Yeon Wei, David Brooks, Gage Hills, and Carole-Jean Wu. 2025. CORDOBA: Carbon-Efficient Opti- mization Framework for Computing Systems. In2025 IEEE Interna- tional Symposium on High Performance Computer Architecture (HPCA). 1289–1303. doi:10.1109/HPCA61900.2025.00098

    work page doi:10.1109/hpca61900.2025.00098 2025

  29. [29]

    2014.Using radio frequency identification (RFID) for moni- toring trees in the forest: State-of-the-technology investigation

    Rey Farve. 2014.Using radio frequency identification (RFID) for moni- toring trees in the forest: State-of-the-technology investigation. United States Department of Agriculture, Forest Service, National Technol- ogy

    work page 2014

  30. [30]

    2014.Using Radio Frequency Identification (RFID) for Monitoring Trees in the Forest: State-of-the-technology Investigation

    Rey Farve. 2014.Using Radio Frequency Identification (RFID) for Monitoring Trees in the Forest: State-of-the-technology Investigation. United States Department of Agriculture, Forest Service, National Technology

    work page 2014

  31. [31]

    Ahmet Feyzioglu and Yavuz Selim Taspinar. 2023. Beef Quality Classification with Reduced E-Nose Data Features According to Beef Cut Types.Sensors23, 4 (Jan. 2023), 2222. doi:10.3390/s23042222

    work page doi:10.3390/s23042222 2023

  32. [32]

    Carloni, and Kenneth Shepard

    Yatin Gilhotra, Henry Overhauser, Heyu Yin, Eric Pollmann, Guy Eichler, Andrew Cheng, Taesung Jung, Nanyu Zeng, Luca P. Carloni, and Kenneth Shepard. 2024. A Wireless Subdural Optical Cortical Interface Device with 768 Co-Packaged Micro-LEDs for Fluorescence Imaging and Optogenetic Stimulation. In2024 IEEE Custom Inte- grated Circuits Conference (CICC). 1...

    work page doi:10.1109/cicc60959.2024 2024

  33. [33]

    Government of India. 2022. Open Government Data (OGD) Platform India. https://data.gov.in

    work page 2022

  34. [34]

    Danielle Grey-Stewart, David Kong, Mariam Elgamal, Georgios Kyr- iazidis, Jalil Morris, and Gage Hills. 2025. Quantifying Trade-Offs in Power, Performance, Area, and Total Carbon Footprint of Future Three-Dimensional Integrated Computing Systems. In2025 Design, Automation & Test in Europe Conference (DATE). IEEE, 1–7

    work page 2025

  35. [35]

    Lee, David Brooks, and Carole-Jean Wu

    Udit Gupta, Mariam Elgamal, Gage Hills, Gu-Yeon Wei, Hsien-Hsin S. Lee, David Brooks, and Carole-Jean Wu. 2022. ACT: Designing Sustainable Computer Systems with an Architectural Carbon Mod- eling Tool. InProceedings of the 49th Annual International Sympo- sium on Computer Architecture. ACM, New York New York, 784–799. doi:10.1145/3470496.3527408

    work page doi:10.1145/3470496.3527408 2022

  36. [36]

    Lee, Gu-Yeon Wei, David Brooks, and Carole-Jean Wu

    Udit Gupta, Young Geun Kim, Sylvia Lee, Jordan Tse, Hsien-Hsin S. Lee, Gu-Yeon Wei, David Brooks, and Carole-Jean Wu. 2021. Chas- ing Carbon: The Elusive Environmental Footprint of Computing. In 2021 IEEE International Symposium on High-Performance Computer Architecture (HPCA). 854–867. doi:10.1109/HPCA51647.2021.00076

    work page doi:10.1109/hpca51647.2021.00076 2021

  37. [37]

    Tom Hackenberg. 2022. MCU prices to remain higher for sev- eral years – Microcontroler (MCU) Quarterly Market Monitor. https://www.yolegroup.com/strategy-insights/mcu-prices-to- remain-higher-for-several-years-microcontroler-mcu-quarterly- market-monitor/

    work page 2022

  38. [38]

    Ramsey Hamade, Raghid Al Ayache, Makram Bou Ghanem, Sleiman El Masri, and Ali Ammouri. 2020. Life Cycle Analysis of AA Alkaline Batteries.Procedia Manufacturing43 (Jan. 2020), 415–422. doi:10.1016/j.promfg.2020.02.193

    work page doi:10.1016/j.promfg.2020.02.193 2020

  39. [39]

    1986.The commercial storage of fruits, vegetables, and florist and nursery stocks

    Robert E Hardenburg, Alley E Watada, and Chien Yi Wang. 1986.The commercial storage of fruits, vegetables, and florist and nursery stocks. Number 66. US Department of Agriculture, Agricultural Research Service

    work page 1986

  40. [40]

    Mark Hempstead, Nikhil Tripathi, Patrick Mauro, Gu-Yeon Wei, and David Brooks. 2005. An ultra low power system architecture for sensor network applications. In32nd International Symposium on Computer Architecture (ISCA’05). IEEE, 208–219

    work page 2005

  41. [41]

    Hilty and Bernard Aebischer

    Lorenz M. Hilty and Bernard Aebischer. 2015. ICT for Sustainability: An Emerging Research Field. InICT Innovations for Sustainability, Lorenz M. Hilty and Bernard Aebischer (Eds.). Springer International Publishing, Cham, 3–36. doi:10.1007/978-3-319-09228-7_1

    work page doi:10.1007/978-3-319-09228-7_1 2015

  42. [42]

    Mark A. J. Huijbregts, Zoran J. N. Steinmann, Pieter M. F. Elshout, Gea Stam, Francesca Verones, Marisa Vieira, Michiel Zijp, Anne Hollander, and Rosalie van Zelm. 2017. ReCiPe2016: A Harmonised Life Cycle Impact Assessment Method at Midpoint and Endpoint Level.The International Journal of Life Cycle Assessment22, 2 (Feb. 2017), 138–147. doi:10.1007/s1136...

    work page doi:10.1007/s11367-016-1246-y 2017

  43. [43]

    IC Insights. 2022. Microcontrollers Get a Lift from Automotive After 2021 Rebound.https://www.icinsights.com/news/bulletins/ microcontrollers-get-a-lift-from-automotive-after-2021-rebound/. Cites 2021 average selling price (ASP) of MCUs as $0.64, based on the Q1 Update to the 2022 McClean Report

    work page 2022

  44. [44]

    Fortune Business Insights. 2025. Smart Clothing Market Size. [On- line]. Available:https://www.fortunebusinessinsights.com/smart- clothing-market-104648

    work page 2025

  45. [45]

    Fortune Business Insights. 2025. Wound Care Devices Market Size. [Online]. Available:https://www.fortunebusinessinsights.com/ industry-reports/wound-care-devices-market-100388

    work page 2025

  46. [46]

    Stanley Jevons

    W. Stanley Jevons. 1998. The Coal Question. InThe Economics of Population. Routledge

    work page 1998

  47. [47]

    Yuanwen Jiang, Artem A Trotsyuk, and Zhenan Bao. 2023. Heal- ing chronic wounds with a wireless smart bandage with integrated sensors and stimulators. 622–623 pages

    work page 2023

  48. [48]

    Antoine Jonquais and Florian Krempl. 2019. Predicting shipping time with machine learning. (2019)

    work page 2019

  49. [49]

    Aydin Kaya, Ali Seydi Keçeli, Cagatay Catal, and Bedir Tekiner- dogan. 2020. Sensor Failure Tolerable Machine Learning-Based Food Quality Prediction Model.Sensors20, 11 (Jan. 2020), 3173. doi:10.3390/s20113173

    work page doi:10.3390/s20113173 2020

  50. [50]

    Shadman Khan, Zeinab Hosseinidoust, Yingfu Li, Carlos DM Filipe, and Tohid F Didar. 2024. Smart food packaging commercialization. Nature Reviews Bioengineering(2024), 1–3

    work page 2024

  51. [51]

    Joshua Kim, En-Fan Chou, Jamie Le, Sabrina Wong, Michael Chu, and Michelle Khine. 2019. Soft wearable pressure sensors for beat- to-beat blood pressure monitoring.Advanced healthcare materials8, 13 (2019), 1900109

    work page 2019

  52. [52]

    Olof Kindgren. 2025. Olofk/Serv: SERV - The SErial RISC-V CPU. https://github.com/olofk/serv

    work page 2025

  53. [53]

    Chrysanthis, and Alex K

    Donald Kline, Nikolas Parshook, Xiaoyu Ge, Erik Brunvand, Rami Melhem, Panos K. Chrysanthis, and Alex K. Jones. 2019. GreenChip: A Tool for Evaluating Holistic Sustainability of Modern Computing Systems.Sustainable Computing: Informatics and Systems22 (June 2019), 322–332. doi:10.1016/j.suscom.2017.10.001

    work page doi:10.1016/j.suscom.2017.10.001 2019

  54. [54]

    Erik Klint, Gregory Peters, and Lars-Olof Johansson. 2024. Pro- Environmental Behaviour Is Undermined by Disgust Sensitivity: The Case of Excessive Laundering.PLOS ONE19, 6 (June 2024), e0302625. doi:10.1371/journal.pone.0302625

    work page doi:10.1371/journal.pone.0302625 2024

  55. [55]

    David Kong, Shvetank Prakash, Jedrzej Kufel, Georgios Kyriazidis, Yasmine Omri, David Verity, Emre Ozer, Vijay Janapa Reddi, and Gage Hills. 2025. 333-eDRAM – 3T Embedded DRAM Leveraging Monolithic 3D Integration of 3 Transistor Types: IGZO, Carbon Nan- otube and Silicon FETs. In2025 62nd ACM/IEEE Design Automation Conference (DAC). IEEE, 1–7

    work page 2025

  56. [56]

    Kumar and B

    K. Kumar and B. P. Pande. 2023. Air Pollution Prediction with Ma- chine Learning: A Case Study of Indian Cities.International Journal of Environmental Science and Technology20, 5 (May 2023), 5333–5348. doi:10.1007/s13762-022-04241-5

    work page doi:10.1007/s13762-022-04241-5 2023

  57. [57]

    Manoj Kumar and Avinash Puri. 2012. A Review of Permissible Limits of Drinking Water.Indian Journal of Occupational and Environmental Medicine16, 1 (2012), 40–44. doi:10.4103/0019-5278.99696

    work page doi:10.4103/0019-5278.99696 2012

  58. [58]

    Kirsi Laitala and Ingun Grimstad Klepp. 2020. What affects garment lifespans? International clothing practices based on a wardrobe sur- vey in China, Germany, Japan, the UK, and the USA.Sustainability 12, 21 (2020), 9151. 13

    work page 2020

  59. [59]

    Emily Laver. 2024. Solid State Battery Life Cycle Analysis: Comparing CO2 Impact of Incumbent Li-Ion Chemistries vs. Emergent Solid State Technology

    work page 2024

  60. [60]

    Baolin Li, Rohan Basu Roy, Daniel Wang, Siddharth Samsi, Vijay Gadepally, and Devesh Tiwari. 2023. Toward Sustainable HPC: Car- bon Footprint Estimation and Environmental Implications of HPC Systems. InProceedings of the International Conference for High Per- formance Computing, Networking, Storage and Analysis. ACM, Denver CO USA, 1–15. doi:10.1145/35817...

    work page doi:10.1145/3581784.3607035 2023

  61. [61]

    Another look at measures of forecast accuracy,

    Dominik Linz, Monika Gawalko, Konstanze Betz, Jeroen M. Hendriks, Gregory Y.H. Lip, Nicklas Vinter, Yutao Guo, and Søren Johnsen. 2024. Atrial Fibrillation: Epidemiology, Screening and Digital Health.The Lancet Regional Health - Europe37 (Feb. 2024), 100786. doi:10.1016/j. lanepe.2023.100786

    work page doi:10.1016/j 2024

  62. [62]

    Jaqueline Litardo, David Gomez, Andrea Boero, Ruben Hidalgo-Leon, Guillermo Soriano, and Angel D Ramirez. 2023. Air-conditioning life cycle assessment research: a review of the methodology, envi- ronmental impacts, and areas of future improvement.Energy and Buildings296 (2023), 113415

    work page 2023

  63. [63]

    Brandon Lucia, Vignesh Balaji, Alexei Colin, Kiwan Maeng, and Emily Ruppel. 2017. Intermittent computing: Challenges and opportunities. 2nd Summit on Advances in Programming Languages (SNAPL 2017) (2017), 8–1

    work page 2017

  64. [64]

    Markets and Markets. 2024. Environmental Monitoring Market Size. [Online]. Available:https://www.marketsandmarkets.com/Market- Reports/environmental-monitoring-market-216846315.html

    work page 2024

  65. [65]

    Markets and Markets. 2024. Smart Agriculture Market Size. [Online]. Available:https://www.marketsandmarkets.com/Market-Reports/ smart-agriculture-market-239736790.html

    work page 2024

  66. [66]

    Ross E Mitchell and Travas A Young. 2023. How Many People Use Sign Language? A National Health Survey-Based Estimate.The Journal of Deaf Studies and Deaf Education28, 1 (Jan. 2023), 1–6. doi:10.1093/deafed/enac031

    work page doi:10.1093/deafed/enac031 2023

  67. [67]

    Arias, and Jan M

    Ali Moin, Andy Zhou, Abbas Rahimi, Alisha Menon, Simone Benatti, George Alexandrov, Senam Tamakloe, Jonathan Ting, Natasha Ya- mamoto, Yasser Khan, Fred Burghardt, Luca Benini, Ana C. Arias, and Jan M. Rabaey. 2021. A Wearable Biosensing System with In-Sensor Adaptive Machine Learning for Hand Gesture Recognition.Nature Electronics4, 1 (Jan. 2021), 54–63....

    work page doi:10.1038/s41928-020-00510-8 2021

  68. [68]

    2013.Water for Food Water for Life: A Comprehensive Assessment of Water Management in Agriculture

    David Molden. 2013.Water for Food Water for Life: A Comprehensive Assessment of Water Management in Agriculture. Routledge

    work page 2013

  69. [69]

    Pooria Mostafalu, William Lenk, Mehmet R Dokmeci, Babak Ziaie, Ali Khademhosseini, and Sameer R Sonkusale. 2015. Wireless flexible smart bandage for continuous monitoring of wound oxygenation. IEEE Transactions on biomedical circuits and systems9, 5 (2015), 670– 677

    work page 2015

  70. [70]

    Muhammad Husnain Mubarik, Dennis D Weller, Nathaniel Bleier, Matthew Tomei, Jasmin Aghassi-Hagmann, Mehdi B Tahoori, and Rakesh Kumar. 2020. Printed machine learning classifiers. In2020 53rd Annual IEEE/ACM International Symposium on Microarchitecture (MICRO). IEEE, 73–87

    work page 2020

  71. [71]

    Arokia Nathan, Arman Ahnood, Matthew T. Cole, Sungsik Lee, Yuji Suzuki, Pritesh Hiralal, Francesco Bonaccorso, Tawfique Hasan, Luis Garcia-Gancedo, Andriy Dyadyusha, Samiul Haque, Piers Andrew, Stephan Hofmann, James Moultrie, Daping Chu, Andrew J. Fle- witt, Andrea C. Ferrari, Michael J. Kelly, John Robertson, Gehan A. J. Amaratunga, and William I. Milne...

    work page doi:10.1109/jproc.2012.2190168 2012

  72. [72]

    United Nations. 2018. 2018 Revision of World Urbaniza- tion Prospects. https://www.un.org/en/desa/2018-revision-world- urbanization-prospects

    work page 2018

  73. [73]

    Emre Ozer, Jedrzej Kufel, John Biggs, Gavin Brown, James Myers, Anjit Rana, Antony Sou, and Catherine Ramsdale. 2019. Bespoke Ma- chine Learning Processor Development Framework on Flexible Sub- strates. In2019 IEEE International Conference on Flexible and Printable Sensors and Systems (FLEPS). 1–3. doi:10.1109/FLEPS.2019.8792285

    work page doi:10.1109/fleps.2019.8792285 2019

  74. [74]

    Emre Ozer, Jedrzej Kufel, John Biggs, James Myers, Charles Reynolds, Gavin Brown, Anjit Rana, Antony Sou, Catherine Ramsdale, and Scott White. 2020. Binary Neural Network as a Flexible Integrated Circuit for Odour Classification. In2020 IEEE International Conference on Flexible and Printable Sensors and Systems (FLEPS). 1–4. doi:10. 1109/FLEPS49123.2020.9239529

    work page arXiv 2020

  75. [75]

    Emre Ozer, Jedrzej Kufel, John Biggs, Anjit Rana, Francisco J Ro- driguez, Thomas Lee-Clark, Antony Sou, Catherine Ramsdale, Scott White, Suresh Kumar Garlapati, et al . 2023. Malodour classifica- tion with low-cost flexible electronics.Nature Communications14, 1 (2023), 777

    work page 2023

  76. [77]

    Emre Ozer, Jedrzej Kufel, Shvetank Prakash, Alireza Raisiardali, Olof Kindgren, Ronald Wong, Nelson Ng, Damien Jausseran, Feras Alkhalil, David Kong, et al. 2024. Bendable non-silicon RISC-V mi- croprocessor.Nature634, 8033 (2024), 341–346

    work page 2024

  77. [78]

    Emre Ozer, Jedrzej Kufel, Elena Punskaya, Igor Drokov, Ari Ercole, Hiral Radia, Thom Smith, John Biggs, Francisco Rodriguez, Grace Aligada, and Adam Rearden. 2024. A Custom-designed Atrial Fib- rillation Detection hardware on a Flexible Substrate. In2024 IEEE International Conference on Flexible and Printable Sensors and Systems (FLEPS). 1–4

    work page 2024

  78. [79]

    Harshil Patel. 2020. INTELLIGENT IRRIGATION SYSTEM. https://www.kaggle.com/datasets/harshilpatel355/autoirrigationdata

    work page 2020

  79. [80]

    Rylie E. O. Pelton, Clare E. Kazanski, Shamitha Keerthi, Kelly A. Racette, Sasha Gennet, Nathaniel Springer, Eugene Yacobson, Michael Wironen, Deepak Ray, Kris Johnson, and Jennifer Schmitt. 2024. Greenhouse Gas Emissions in US Beef Production Can Be Reduced by up to 30% with the Adoption of Selected Mitigation Measures.Nature Food5, 9 (Sept. 2024), 787–7...

    work page doi:10.1038/s43016-024-01031-9 2024

  80. [81]

    PhysioNet. 2000. MIT-BIH Atrial Fibrillation Database.https:// physionet.org/content/afdb/1.0.0

    work page 2000

Showing first 80 references.