Special Topic on Non-traditional Computing for Energy Efficiency

Guest Editors

Kerem Y. Camsari, University of California, Santa Barbara,

Supriyo Datta, Purdue University,

Editor-in-Chief

Azad Naeemi, Georgia Institute of Technology,

Aim and Scope

Computing now consumes a significant portion of global electricity production, with AI and data center workloads projected to reach larger scales if current trends continue. The continued success of computing technologies depends on dramatic improvements in energy efficiency across devices, circuits, and systems. While CMOS scaling has powered decades of progress, it now faces fundamental limits in power and density. It is increasingly mismatched to the needs of data-intensive applications such as artificial intelligence, optimization, and edge computing. This has motivated a broad search for non-traditional computing approaches that harness physical dynamics, stochasticity, memory-in-logic integration, and new materials to move beyond the von Neumann paradigm.

Non-traditional computing spans a wide spectrum: from new devices such as spintronic, ferroelectric, photonic, superconducting, and phase-transition elements, to oscillator and probabilistic networks, to hybrid CMOS + X architectures that

co-design physics and algorithms. These approaches promise energy savings through new forms of computation rooted in natural dynamics. Realizing this potential requires coordinated advances across the technology stack. Their impact depends on bridging scales from material and device physics to compact models to system-level demonstrations and applications in machine learning, optimization, and scientific computing.

This Special Topic of the IEEE Journal on Exploratory Solid-State Computational Devices and Circuits (JxCDC) will highlight advances in energy-efficient non-traditional computing that demonstrate the potential for significant power consumption reductions across devices, circuits, architectures, and algorithms. We invite contributions showing physical realizations, modeling, simulation, and co-design strategies that push the frontiers of exploratory computing beyond conventional computing paradigms.

Topics of Interest

Topics of interest include, but are not limited to:

Device primitives: spintronic, ferroelectric, photonic, superconducting, resistive, correlated-electron, and phase-transition devices for non-traditional computing
Stochastic and dynamical elements: probabilistic bits, noisy switching devices, oscillator and neuromorphic circuits, networks exploiting natural dynamics
Circuits and architectures: neuromorphic processors, memory-in-logic, hybrid CMOS + X systems (e.g., CMOS-spintronic, CMOS-memristive), distributed and parallel architectures such as Ising and oscillator networks
Modeling and simulation: first principles and physics-based models, compact device models, multiscale frameworks linking materials to circuits, benchmarking methodologies
Algorithm-hardware co-design: physics-inspired algorithms leveraging stochasticity and dynamics, hybrid probabilistic-deterministic implementations, and frameworks for large-scale optimization and generative AI
Quantum-inspired approaches: tensor networks, variational algorithms implemented classically, and quantum annealing-inspired architectures
Experimental validations: prototype demonstrations, proof-of-concept systems, comparative studies with CMOS baselines
Applications: energy-efficient hardware for generative AI, machine learning, combinatorial optimization, scientific computing
Cross-cutting issues: variability and noise as computational resources, reliability and scalability, heterogeneous integration with CMOS, and energy-efficiency metrics across devices and systems

Submission Guidelines

Submit your paper through the JXCDC submission site.

Deadlines

Open for Submission: October 15, 2025
Submission Deadline: January 15, 2026
First Notification: February 15, 2026
Revision Submission: March 1, 2026
Final Decision: March 15, 2026
Online Special Topic Publication: April 1, 2026

Papers submitted earlier than the submission deadline will be reviewed upon submission and will be published earlier than the timeline listed above.