TEL-RAM: A Specialized Memory Architecture for Energy-Efficient Acceleration via Thermo-Electric Sensing

This Version(1.1) is Repaired For Version 1.0 This paper proposes TEL-RAM (Thermo-Electric Loop RAM), a memory architecture that de-couples storage, readout, and maintenance functions. Using a storage coil, passive electromagnetic sensing, and thermal-based replenishment control, TEL-RAM achieves ultra-low read energy (˜0.1–1pJ), minimal write energy (<0.01 pJ), and extremely low standby power (˜0.017 pW/bit). The cell requires multiple switches and a comparator, resulting in a larger area (˜1.0–2.0 μm2 at 28nm) compared to DRAM, positioning it as a specialized accelerator memory rather than a general-purpose DRAM replacement. Due to its cell area (˜1.0–2.0 μm2/bit at 28nm), TEL-RAM is not suited for large-scale mainmemory or for enabling “space-for-time” trade-offs at capacity. Instead, it functions as a specialized,low-density accelerator memory for read-intensive, batch-write-dominant, and energy-constrained workloads. A complete energy model, latency analysis, thermal time constant verification, SNR estimation, and area breakdown are provided. The architecture is compared against SRAM, DRAM,MRAM, and ReRAM, with clear positioning for AI inference, edge computing, and 2.5D heterogeneous integration.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-08-21
DOI
https://doi.org/10.5281/zenodo.21735491
Primary Topic
Advanced Memory and Neural Computing
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

TEL-RAM: A Specialized Memory Architecture for Energy-Efficient Acceleration via Thermo-Electric Sensing

Li-Kuang Yin
Zenodo (CERN European Organization for Nuclear Research)
Advanced Memory and Neural Computing
preprint

TEL-RAM: A Specialized Memory Architecture for Energy-Efficient Acceleration via Thermo-Electric Sensing

Li-Kuang Yin
preprint en

Abstract

This Version(1.1) is Repaired For Version 1.0 This paper proposes TEL-RAM (Thermo-Electric Loop RAM), a memory architecture that de-couples storage, readout, and maintenance functions. Using a storage coil, passive electromagnetic sensing, and thermal-based replenishment control, TEL-RAM achieves ultra-low read energy (˜0.1–1pJ), minimal write energy (<0.01 pJ), and extremely low standby power (˜0.017 pW/bit). The cell requires multiple switches and a comparator, resulting in a larger area (˜1.0–2.0 μm2 at 28nm) compared to DRAM, positioning it as a specialized accelerator memory rather than a general-purpose DRAM replacement. Due to its cell area (˜1.0–2.0 μm2/bit at 28nm), TEL-RAM is not suited for large-scale mainmemory or for enabling “space-for-time” trade-offs at capacity. Instead, it functions as a specialized,low-density accelerator memory for read-intensive, batch-write-dominant, and energy-constrained workloads. A complete energy model, latency analysis, thermal time constant verification, SNR estimation, and area breakdown are provided. The architecture is compared against SRAM, DRAM,MRAM, and ReRAM, with clear positioning for AI inference, edge computing, and 2.5D heterogeneous integration.

Zenodo (CERN European Organization for Nuclear Research)
Affordable and clean energy
Advanced Memory and Neural Computing
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.