arXiv subjects

Adam Teman

Publications and source records attributed to Adam Teman.

10 recordsLinked to original sources

arXiv2026

EULER-ADAS: Energy-Efficient & SIMD-Unified Logarithmic-Posit Engine for Precision-Reconfigurable Approximate ADAS Acceleration

Advanced driver-assistance systems (ADAS) require neural compute engines that deliver low-latency inference under strict power and area constraints. Posit arithmetic is attractive for such accelerators because it provides high numerical fidelity at low precision, but its variable-length regime encoding increases encode/decode cost and exposes the datapath to large regime-field fault effects. This paper presents EULER-ADAS, a SIMD-enabled logarithmic bounded-Posit neural compute engine for energyefficient and reliability-aware ADAS acceleration. The proposed datapath combines bounded-regime Posit representation, stageadaptive logarithmic mantissa multiplication with bit truncation, and a SIMD-shared quire accumulation path supporting Posit- (8,0), Posit-(16,1), and Posit-(32,2) execution. The unified architecture enables 4xPosit-8, 2xPosit-16, or 1xPosit-32 operation without duplicating precision-specific hardware. FPGA implementation shows that the proposed configurations reduce LUT count by up to 41.4%, delay by up to 76.1%, and power by up to 71.9% relative to exact Posit neural compute engines, while achieving up to 10x lower energy-delay product than radix-4 Booth-based Posit multipliers. In 28-nm CMOS, the bounded variants occupy 0.013-0.016 mm2 , consume 19.8-22.1 mW, and operate at up to 1.84 GHz. Application-level evaluation across image-classification, ADAS, and edge-inference workloads shows that the evaluated Posit-16 and Posit-32 configurations remain within about 1.5 percentage points of FP32 accuracy. A TinyYOLOv3 prototype on Pynq-Z2 achieves 78 ms latency at 0.29 W and 22.6 mJ/frame, demonstrating the suitability of EULERADAS for low-power real-time ADAS inference.

cs.AR↗

arXiv2026

CARMEN: CORDIC-Accelerated Resource-Efficient Multi-Precision Inference Engine for Deep Learning

This paper presents CARMEN, a runtime-adaptive, CORDIC-accelerated multi-precision vector engine for resource-efficient deep learning inference. The key insight is that CORDIC iteration depth directly governs computational accuracy, enabling dynamic switching between approximate and accurate execution modes without hardware modification. The architecture integrates a low-resource iterative CORDIC-based MAC unit with a time-multiplexed multi-activation function block, supporting flexible 8/16-bit precision and high hardware utilization. ASIC implementation in 28 nm CMOS achieves up to 33% reduction in computation cycles and 21% power savings per MAC stage; a 256-PE configuration delivers 4.83 TOPS/mm2 compute density and 11.67 TOPS/W energy efficiency. FPGA deployment on PynqZ2 validates 154.6 ms latency at 0.43 W for real-time object detection.

cs.AR↗

arXiv2026

SPADE: A SIMD Posit-enabled compute engine for Accelerating DNN Efficiency

The growing demand for edge-AI systems requires arithmetic units that balance numerical precision, energy efficiency, and compact hardware while supporting diverse formats. Posit arithmetic offers advantages over floating- and fixed-point representations through its tapered precision, wide dynamic range, and improved numerical robustness. This work presents SPADE, a unified multi-precision SIMD Posit-based multiplyaccumulate (MAC) architecture supporting Posit (8,0), Posit (16,1), and Posit (32,2) within a single framework. Unlike prior single-precision or floating/fixed-point SIMD MACs, SPADE introduces a regime-aware, lane-fused SIMD Posit datapath that hierarchically reuses Posit-specific submodules (LOD, complementor, shifter, and multiplier) across 8/16/32-bit precisions without datapath replication. FPGA implementation on a Xilinx Virtex-7 shows 45.13% LUT and 80% slice reduction for Posit (8,0), and up to 28.44% and 17.47% improvement for Posit (16,1) and Posit (32,2) over prior work, with only 6.9% LUT and 14.9% register overhead for multi-precision support. ASIC results across TSMC nodes achieve 1.38 GHz at 6.1 mW (28 nm). Evaluation on MNIST, CIFAR-10/100, and alphabet datasets confirms competitive inference accuracy.

cs.AR↗

arXiv2026

PiC-BNN: A 128-kbit 65 nm Processing-in-CAM-Based End-to-End Binary Neural Network Accelerator

Binary Neural Networks (BNNs), where weights and activations are constrained to binary values (+1, -1), are a highly efficient alternative to traditional neural networks. Unfortunately, typical BNNs, while binarizing linear layers (matrix-vector multiplication), still implement other network layers (batch normalization, softmax, output layer, and sometimes the input layer of a convolutional neural network) in full precision. This limits the area and energy benefits and requires architectural support for full precision operations. We propose PiC-BNN, a true end-to-end binary in-approximate search (Hamming distance tolerant) Content Addressable Memory based BNN accelerator. PiC-BNN is designed and manufactured in a commercial 65nm process. PiC-BNN uses Hamming distance tolerance to apply the law of large numbers to enable accurate classification without implementing full precision operations. PiC-BNN achieves baseline software accuracy (95.2%) on the MNIST dataset and 93.5% on the Hand Gesture (HG) dataset, a throughput of 560K inferences/s, and presents a power efficiency of 703M inferences/s/W when implementing a binary MLP model for MNIST/HG dataset classification.

cs.AR↗

arXiv2025

Bhasha-Rupantarika: Algorithm-Hardware Co-design approach for Multilingual Neural Machine Translation

This paper introduces Bhasha-Rupantarika, a light and efficient multilingual translation system tailored through algorithm-hardware codesign for resource-limited settings. The method investigates model deployment at sub-octet precision levels (FP8, INT8, INT4, and FP4), with experimental results indicating a 4.1x reduction in model size (FP4) and a 4.2x speedup in inference speed, which correlates with an increased throughput of 66 tokens/s (improvement by 4.8x). This underscores the importance of ultra-low precision quantization for real-time deployment in IoT devices using FPGA accelerators, achieving performance on par with expectations. Our evaluation covers bidirectional translation between Indian and international languages, showcasing its adaptability in low-resource linguistic contexts. The FPGA deployment demonstrated a 1.96x reduction in LUTs and a 1.65x decrease in FFs, resulting in a 2.2x enhancement in throughput compared to OPU and a 4.6x enhancement compared to HPTA. Overall, the evaluation provides a viable solution based on quantisation-aware translation along with hardware efficiency suitable for deployable multilingual AI systems. The entire codes [https://github.com/mukullokhande99/Bhasha-Rupantarika/] and dataset for reproducibility are publicly available, facilitating rapid integration and further development by researchers.

cs.AR↗

arXiv2025

Retrospective: A CORDIC Based Configurable Activation Function for NN Applications

A CORDIC-based configuration for the design of Activation Functions (AF) was previously suggested to accelerate ASIC hardware design for resource-constrained systems by providing functional reconfigurability. Since its introduction, this new approach for neural network acceleration has gained widespread popularity, influencing numerous designs for activation functions in both academic and commercial AI processors. In this retrospective analysis, we explore the foundational aspects of this initiative, summarize key developments over recent years, and introduce the DA-VINCI AF tailored for the evolving needs of AI applications. This new generation of dynamically configurable and precision-adjustable activation function cores promise greater adaptability for a range of activation functions in AI workloads, including Swish, SoftMax, SeLU, and GeLU, utilizing the Shift-and-Add CORDIC technique. The previously presented design has been optimized for MAC, Sigmoid, and Tanh functionalities and incorporated into ReLU AFs, culminating in an accumulative NEURIC compute unit. These enhancements position NEURIC as a fundamental component in the resource-efficient vector engine for the realization of AI accelerators that focus on DNNs, RNNs/LSTMs, and Transformers, achieving a quality of results (QoR) of 98.5%.

cs.AR↗

arXiv2025

CORDIC Is All You Need

Artificial intelligence necessitates adaptable hardware accelerators for efficient high-throughput million operations. We present pipelined architecture with CORDIC block for linear MAC computations and nonlinear iterative Activation Functions (AF) such as $tanh$, $sigmoid$, and $softmax$. This approach focuses on a Reconfigurable Processing Engine (RPE) based systolic array, with 40\% pruning rate, enhanced throughput up to 4.64$\times$, and reduction in power and area by 5.02 $\times$ and 4.06 $\times$ at CMOS 28 nm, with minor accuracy loss. FPGA implementation achieves a reduction of up to 2.5 $\times$ resource savings and 3 $\times$ power compared to prior works. The Systolic CORDIC engine for Reconfigurability and Enhanced throughput (SYCore) deploys an output stationary dataflow with the CAESAR control engine for diverse AI workloads such as Transformers, RNNs/LSTMs, and DNNs for applications like image detection, LLMs, and speech recognition. The energy-efficient and flexible approach extends the enhanced approach for edge AI accelerators supporting emerging workloads.

cs.AR↗

arXiv2022

Adjusting Thermal Stability in Double-Barrier MTJ for Energy Improvement in Cryogenic STT-MRAMs

This paper investigates the impact of thermal stability relaxation in double-barrier magnetic tunnel junctions (DMTJs) for energy-efficient spin-transfer torque magnetic random access memories (STT-MRAMs) operating at the liquid nitrogen boiling point (77K). Our study is carried out through a macrospin-based Verilog-A compact model of DMTJ, along with a 65nm commercial process design kit (PDK) calibrated down to 77K under silicon measurements. Comprehensive bitcell-level electrical characterization is used to estimate the energy/latency per operation and leakage power at the memory architecture-level. As a main result of our analysis, we show that energy-efficient small-to-large embedded memories can be obtained by significantly relaxing the non-volatility requirement of DMTJ devices at room temperature (i.e., by reducing the cross-section area), while maintaining the typical 10-years retention time at cryogenic temperatures. This makes DMTJ-based STT-MRAM operating at 77K more energy-efficient than six-transistors static random-access memory (6T-SRAM) under both read and write accesses (-56% and -37% on average, respectively). Obtained results thus prove that DMTJ-based STT-MRAM with relaxed retention time is a promising alternative for the realization of reliable and energy-efficient embedded memories operating at cryogenic temperatures.

cs.ET↗

arXiv2021

Hamming Distance Tolerant Content-Addressable Memory (HD-CAM) for Approximate Matching Applications

We propose a novel Hamming distance tolerant content-addressable memory (HD-CAM) for energy-efficient in memory approximate matching applications. HD-CAM implements approximate search using matchline charge redistribution rather than its rise or fall time, frequently employed in state of-the-art solutions. HD-CAM was designed in a 65 nm 1.2 V CMOS technology and evaluated through extensive Monte Carlo simulations. Our analysis shows that HD-CAM supports robust operation under significant process variations and changes in the design parameters, enabling a wide range of mismatch threshold (tolerable Hamming distance) levels and pattern lengths. HD-CAM was functionally evaluated for virus DNA classification, which makes HD-CAM suitable for hardware acceleration of genomic surveillance of viral outbreaks such as Covid-19 pandemics.

cs.AR↗

arXiv2017

A 588 Gbps LDPC Decoder Based on Finite-Alphabet Message Passing

An ultra-high throughput low-density parity check (LDPC) decoder with an unrolled full-parallel architecture is proposed, which achieves the highest decoding throughput compared to previously reported LDPC decoders in the literature. The decoder benefits from a serial message-transfer approach between the decoding stages to alleviate the well-known routing congestion problem in parallel LDPC decoders. Furthermore, a finite-alphabet message passing algorithm is employed to replace the variable node update rule of the standard min-sum decoder with look-up tables, which are designed in a way that maximizes the mutual information between decoding messages. The proposed algorithm results in an architecture with reduced bit-width messages, leading to a significantly higher decoding throughput and to a lower area as compared to a min-sum decoder when serial message-transfer is used. The architecture is placed and routed for the standard min-sum reference decoder and for the proposed finite-alphabet decoder using a custom pseudo-hierarchical backend design strategy to further alleviate routing congestions and to handle the large design. Post-layout results show that the finite-alphabet decoder with the serial message-transfer architecture achieves a throughput as large as 588 Gbps with an area of 16.2 mm$^2$ and dissipates an average power of 22.7 pJ per decoded bit in a 28 nm FD-SOI library. Compared to the reference min-sum decoder, this corresponds to 3.1 times smaller area and 2 times better energy efficiency.

cs.AR↗