arXiv subjects

Giorgos Dimitrakopoulos

Publications and source records attributed to Giorgos Dimitrakopoulos.

At least 19 recordsLinked to original sources

arXiv2026

High-Performance NTT Accelerators for PQC leveraging Unified Redundant Arithmetic and Fine-Tuned Microarchitecture

Post-quantum cryptography and privacy-preserving technologies are expected to play a central role in future secure communication systems. Lattice-based PQC schemes such as ML-KEM (CRYSTALS-Kyber) and ML-DSA (CRYSTALS-Dilithium) rely heavily on large-degree polynomial arithmetic, making the Number Theoretic Transform (NTT) a key computational primitive. Although existing hardware accelerators exploit parallelism and pipelining to support both NTT and INTT, their efficiency is often limited by the overhead of modular reduction and correction steps, inverse-transform scaling operations, and suboptimal FPGA implementations. This work addresses these limitations by proposing parallel iterative NTT/INTT accelerators based on optimized unified butterfly units. We introduce a novel redundant number representation that eliminates conditional corrections for both Montgomery modulo multiplication and combined subtract-multiply operations, and integrate inverse-transform scaling into existing arithmetic hardware to avoid dedicated scaling units. Furthermore, we design hierarchical Montgomery multipliers that map efficiently onto FPGA DSP resources, reducing hardware cost while enabling high operating frequencies. FPGA-based experimental results demonstrate higher clock frequencies, reduced execution times, and competitive resource utilization, supporting efficient NTT acceleration for PQC and related privacy-preserving applications.

cs.AR↗

arXiv2026

Low-Cost Multi-Precision Systolic Arrays for Accelerating FHE NTTs on AI ASICs

Fully Homomorphic Encryption (FHE) ensures robust data privacy but suffers from prohibitive computational overhead. Accelerating FHE on AI hardware like Tensor Processing Units (TPUs) is promising, yet fundamentally limited by a precision mismatch: TPUs are optimized for 8-bit arithmetic, whereas FHE and its critical parts such as the Number Theoretic Transform (NTT), demand high precision. Current approaches bridge this gap using matrix decomposition to execute NTT computations on low-precision matrix engines. However, reconstructing the full-precision results requires shift-and-add accumulation that does not match the dataflow of matrix multiplication. This forces offloading full-precision reconstruction from matrix engines to vector processors that disrupts the matrix multiplication dataflow, creating significant performance bottleneck. To resolve this limitation, we propose a minimally modified multi-precision systolic array that performs full-precision output reconstruction natively within the array in sync with low-precision matrix multiplication under a uniform dataflow. Synthesized at 7nm with OpenRoad, our design incurs negligible hardware overhead. Cycle-accurate simulations using SCALE-Sim demonstrate that natively executing NTTs on the proposed architecture achieves at least 1.33x speedup, for transform sizes 2^12 to 2^16 on 128x128 matrix engines, successfully enabling standard AI hardware to support high-precision FHE acceleration.

cs.CR↗

arXiv2026

MIVE: A Minimalist Integer Vector Engine for Softmax LayerNorm and RMSNorm Acceleration

The rapid growth of Large Language Models (LLMs) has intensified the need for specialized hardware accelerators that can satisfy stringent inference latency and power constraints. Although matrix multiplications dominate the overall computational workload, non-linear vector normalization operations, such as LayerNorm, RMSNorm and Softmax can become critical hardware bottlenecks. Existing accelerators typically implement these functions using dedicated hardware blocks, leading to duplicated resources and inefficient silicon utilization. To address this limitation, we propose a Minimalist Integer Vector Engine (MIVE), a programmable architecture capable of executing all three operations within a unified datapath. By exploiting common computational patterns across LayerNorm, RMSNorm and Softmax the proposed vector engine maximizes hardware sharing while reducing implementation overhead. Physical ASIC implementation results show that MIVE provides comprehensive multi-function support while achieving higher area and hardware efficiency than most state-of-the-art standalone accelerators.

cs.AR↗

arXiv2026

MPX: A Unified Systolic Array for Matrix and Polynomial Multiplication

Polynomial multiplication is a fundamental kernel in Fully Homomorphic Encryption (FHE) and post-quantum cryptography (PQC) and is commonly accelerated through Number Theoretic Transforms (NTTs). To avoid the cost of designing dedicated cryptographic accelerators, recent efforts have mapped NTT computations onto existing systolic matrix engines, enabling the reuse of AI hardware for cryptographic workloads. In this work, we take the opposite approach. We observe that the wavefront dataflow of systolic arrays naturally aligns with the accumulation pattern of polynomial multiplication and leverage this correspondence to design MPX, a dual-mode systolic array that supports both matrix multiplication and direct polynomial multiplication within the same hardware fabric. Experimental results show that extending a conventional systolic array with this dual-mode capability requires only 20% additional area and introduces negligible power overhead during matrix-multiplication execution. In polynomial-multiplication mode, MPX achieves more than 1.2x lower latency compared to NTT-based polynomial multiplication on systolic matrix engines.

cs.CR↗

arXiv2026

VQ4SNN: Vector Quantization for Memory-Efficient FPGA Spiking Neural Networks

Spiking Neural Networks (SNNs) offer an energy-efficient paradigm for edge AI, making them attractive for hardware acceleration. However, deploying dense SNNs on FPGAs is constrained by limited on-chip memory for synaptic weight storage. To address this bottleneck, we propose VQ4SNN, a hardware-aware architecture that reduces memory requirements through Vector Quantization (VQ)-based weight sharing. To the best of our knowledge, this is the first application of VQ to pipelined spatial-dataflow SNN accelerators on FPGAs. VQ4SNN replaces conventional weight storage with a two-level memory organization consisting of compact pointers and a shared codebook of quantized weight vectors. The proposed design integrates FPGA-aware memory mapping with analytical VQ parameter selection, enabling efficient deployment on such accelerators while preserving inference accuracy. The experimental results show a reduction of 52-61% in the total number of BRAMs compared to the state-of-the-art uncompressed FPGA SNNs without increasing overall logic utilization.

cs.NE↗

arXiv2026

H-FA: A Hybrid Floating-Point and Logarithmic Approach to Hardware Accelerated FlashAttention

Transformers have significantly advanced AI and machine learning through their powerful attention mechanism. However, computing attention on long sequences can become a computational bottleneck. FlashAttention mitigates this by fusing the softmax and matrix operations into a tiled computation pattern that decouples performance from sequence length. Though designed for GPUs, its simplicity also makes it well suited for direct hardware acceleration. To improve hardware implementation, we compute FlashAttention using a mixture of floating-point and fixed-point logarithm domain representations. Floating-point is used to compute attention scores from query and key matrices, while logarithmic computation simplifies the fused computation of softmax normalization and the multiplication with the value matrix. This transformation, called H-FA, replaces vector-wide floating-point multiplication and division operations by additions and subtractions implemented efficiently with fixed-point arithmetic in the logarithm domain. Exponential function evaluations are effectively omitted and fused with the rest operations, and the final result is directly returned to floating-point arithmetic without any additional hardware overhead. Hardware implementation results at 28nm demonstrate that H-FA achieves a 26.5% reduction in area and a 23.4% reduction in power, on average, compared to FlashAttention parallel hardware architectures built solely with floating-point datapaths, without hindering performance.

cs.AR↗

arXiv2025

Vectorized FlashAttention with Low-cost Exponential Computation in RISC-V Vector Processors

Attention is a core operation in numerous machine learning and artificial intelligence models. This work focuses on the acceleration of attention kernel using FlashAttention algorithm, in vector processors, particularly those based on the RISC-V instruction set architecture (ISA). This work represents the first effort to vectorize FlashAttention, minimizing scalar code and simplifying the computational complexity of evaluating exponentials needed by softmax used in attention. By utilizing a low-cost approximation for exponentials in floating-point arithmetic, we reduce the cost of computing the exponential function without the need to extend baseline vector ISA with new custom instructions. Also, appropriate tiling strategies are explored with the goal to improve memory locality. Experimental results highlight the scalability of our approach, demonstrating significant performance gains with the vectorized implementations when processing attention layers in practical applications.

cs.LG↗

arXiv2025

Custom Algorithm-based Fault Tolerance for Attention Layers in Transformers

Transformers and large language models (LLMs), powered by the attention mechanism, have transformed numerous AI applications, driving the need for specialized hardware accelerators. A major challenge in these accelerators is efficiently detecting errors caused by random hardware faults. Traditional algorithm-based fault tolerance (ABFT) techniques verify individual matrix multiplications but fall short in handling the full attention mechanism, particularly due to intermediate softmax normalization. This work proposes Flash-ABFT, a novel method that computes an online checksum across the entire three-matrix product of query, key and value matrices, of an attention layer, including the softmax operation, with a single check. This approach significantly reduces overhead by eliminating redundant checks while maintaining high fault-detection accuracy. Experimental results demonstrate that Flash-ABFT incurs only 5.3% hardware area overhead and less than 1.9% energy overhead, making it a cost-effective and robust solution for error detection in attention accelerators.

cs.LG↗

arXiv2025

High-Performance Pipelined NTT Accelerators with Homogeneous Digit-Serial Modulo Arithmetic

The Number Theoretic Transform (NTT) is a fundamental operation in privacy-preserving technologies, particularly within fully homomorphic encryption (FHE). The efficiency of NTT computation directly impacts the overall performance of FHE, making hardware acceleration a critical technology that will enable realistic FHE applications. Custom accelerators, in FPGAs or ASICs, offer significant performance advantages due to their ability to exploit massive parallelism and specialized optimizations. However, the operation of NTT over large moduli requires large word-length modulo arithmetic that limits achievable clock frequencies in hardware and increases hardware area costs. To overcome such deficits, digit-serial arithmetic has been explored for modular multiplication and addition independently. The goal of this work is to leverage digit-serial modulo arithmetic combined with appropriate redundant data representation to design modular pipelined NTT accelerators that operate uniformly on arbitrary small digits, without the need for intermediate (de)serialization. The proposed architecture enables high clock frequencies through regular pipelining while maintaining parallelism. Experimental results demonstrate that the proposed approach outperforms state-of-the-art implementations and reduces hardware complexity under equal performance and input-output bandwidth constraints.

cs.AR↗

arXiv2025

Efficient Implementation of RISC-V Vector Permutation Instructions

RISC-V CPUs leverage the RVV (RISC-V Vector) extension to accelerate data-parallel workloads. In addition to arithmetic operations, RVV includes powerful permutation instructions that enable flexible element rearrangement within vector registers --critical for optimizing performance in tasks such as matrix operations and cryptographic computations. However, the diverse control mechanisms of these instructions complicate their execution within a unified datapath while maintaining the fixed-latency requirement of cryptographic accelerators. To address this, we propose a unified microarchitecture capable of executing all RVV permutation instructions efficiently, regardless of their control information structure. This approach minimizes area and hardware costs while ensuring single-cycle execution for short vector machines (up to 256 bits) and enabling efficient pipelining for longer vectors. The proposed design is integrated into an open-source RISC-V vector processor and implemented at 7 nm using the OpenRoad physical synthesis flow. Experimental results validate the efficiency of our unified vector permutation unit, demonstrating that it only incurs 1.5% area overhead to the total vector processor. Furthermore, this area overhead decreases to near-0% as the minimum supported element width for vector permutations increases.

cs.AR↗

arXiv2025

Low-Cost FlashAttention with Fused Exponential and Multiplication Hardware Operators

Attention mechanisms, particularly within Transformer architectures and large language models (LLMs), have revolutionized sequence modeling in machine learning and artificial intelligence applications. To compute attention for increasingly long sequences, specialized accelerators have been proposed to execute key attention steps directly in hardware. Among the various recently proposed architectures, those based on variants of the FlashAttention algorithm, originally designed for GPUs, stand out due to their optimized computation, tiling capabilities, and reduced memory traffic. In this work, we focus on optimizing the kernel of floating-point-based FlashAttention using new hardware operators that fuse the computation of exponentials and vector multiplications, e.g., e^x, V. The proposed ExpMul hardware operators significantly reduce the area and power costs of FlashAttention-based hardware accelerators. When implemented in a 28nm ASIC technology, they achieve improvements of 28.8% in area and 17.6% in power, on average, compared to state-of-the-art hardware architectures with separate exponentials and vector multiplications hardware operators.

cs.AR↗

arXiv2025

FLASH-D: FlashAttention with Hidden Softmax Division

The transformer's attention mechanism has revolutionized AI and machine learning, with its efficient computation being crucial to its performance. However, calculating attention involves matrix operations interspersed with softmax rescaling, which inherently slows down computation and requires processing the entire input sequence. Building on online softmax computation, FlashAttention integrates softmax calculation with matrix arithmetic, enabling tiled computation independent of sequence length. While optimized for GPUs, FlashAttention's simplicity makes it amenable to direct hardware acceleration. This work re-evaluates the core FlashAttention kernel, presenting FLASH-D a mathematically equivalent, yet simplified, formulation that achieves: (a) hiding softmax division within other non-linear function evaluations; (b) inherently numerically stable computation of exponentials, eliminating the need for maximum value subtraction; and (c) a reduction in computational cost without introducing numerical approximations to the FlashAttention kernel. Importantly, the essential FlashAttention properties that facilitate efficient tiled implementation are fully preserved. Hardware implementation results at 28nm demonstrate that this proposed formulation achieves a 22.8% reduction in area and a 20.3% reduction in power, on average, compared to state-of-the-art parallel hardware architectures without any performance penalty.

cs.LG↗

arXiv2025

Periodic Online Testing for Sparse Systolic Tensor Arrays

Modern Machine Learning (ML) applications often benefit from structured sparsity, a technique that efficiently reduces model complexity and simplifies handling of sparse data in hardware. Sparse systolic tensor arrays - specifically designed to accelerate these structured-sparse ML models - play a pivotal role in enabling efficient computations. As ML is increasingly integrated into safety-critical systems, it is of paramount importance to ensure the reliability of these systems. This paper introduces an online error-checking technique capable of detecting and locating permanent faults within sparse systolic tensor arrays before computation begins. The new technique relies on merely four test vectors and exploits the weight values already loaded within the systolic array to comprehensively test the system. Fault-injection campaigns within the gate-level netlist, while executing three well-established Convolutional Neural Networks (CNN), validate the efficiency of the proposed approach, which is shown to achieve very high fault coverage, while incurring minimal performance and area overheads.

cs.AR↗

arXiv2025

Register Dispersion: Reducing the Footprint of the Vector Register File in Vector Engines of Low-Cost RISC-V CPUs

The deployment of Machine Learning (ML) applications at the edge on resource-constrained devices has accentuated the need for efficient ML processing on low-cost processors. While traditional CPUs provide programming flexibility, their general-purpose architecture often lacks the throughput required for complex ML models. The augmentation of a RISC-V processor with a vector unit can provide substantial data-level parallelism. However, increasing the data-level parallelism supported by vector processing would make the Vector Register File (VRF) a major area consumer in ultra low-cost processors, since 32 vector registers are required for RISC-V Vector ISA compliance. This work leverages the insight that many ML vectorized kernels require a small number of active vector registers, and proposes the use of a physically smaller VRF that dynamically caches only the vector registers currently accessed by the application. This approach, called Register Dispersion, maps the architectural vector registers to a smaller set of physical registers. The proposed ISA-compliant VRF is significantly smaller than a full-size VRF and operates like a conventional cache, i.e., it only stores the most recently accessed vector registers. Essential registers remain readily accessible within the compact VRF, while the others are offloaded to the cache/memory sub-system. The compact VRF design is demonstrated to yield substantial area and power savings, as compared to using a full VRF, with no or minimal impact on performance. This effective trade-off renders the inclusion of vector units in low-cost processors feasible and practical.

cs.AR↗

arXiv2025

Optimizing Structured-Sparse Matrix Multiplication in RISC-V Vector Processors

Structured sparsity has been proposed as an efficient way to prune the complexity of Machine Learning (ML) applications and to simplify the handling of sparse data in hardware. Accelerating ML models, whether for training, or inference, heavily relies on matrix multiplications that can be efficiently executed on vector processors, or custom matrix engines. This work aims to integrate the simplicity of structured sparsity into vector execution to speed up the corresponding matrix multiplications. Initially, the implementation of structured-sparse matrix multiplication using the current RISC-V instruction set vector extension is comprehensively explored. Critical parameters that affect performance, such as the impact of data distribution across the scalar and vector register files, data locality, and the effectiveness of loop unrolling are analyzed both qualitatively and quantitatively. Furthermore, it is demonstrated that the addition of a single new instruction would reap even higher performance. The newly proposed instruction is called vindexmac, i.e., vector index-multiply-accumulate. It allows for indirect reads from the vector register file and it reduces the number of instructions executed per matrix multiplication iteration, without introducing additional dependencies that would limit loop unrolling. The proposed new instruction was integrated in a decoupled RISC-V vector processor with negligible hardware cost. Experimental results demonstrate the runtime efficiency and the scalability offered by the introduced optimizations and the new instruction for the execution of state-of-the-art Convolutional Neural Networks. More particularly, the addition of a custom instruction improves runtime by 25% and 33% when compared with highly-optimized vectorized kernels that use only the currently defined RISC-V instructions.

cs.AR↗

arXiv2024

GCN-ABFT: Low-Cost Online Error Checking for Graph Convolutional Networks

Graph convolutional networks (GCNs) are popular for building machine-learning application for graph-structured data. This widespread adoption led to the development of specialized GCN hardware accelerators. In this work, we address a key architectural challenge for GCN accelerators: how to detect errors in GCN computations arising from random hardware faults with the least computation cost. Each GCN layer performs a graph convolution, mathematically equivalent to multiplying three matrices, computed through two separate matrix multiplications. Existing Algorithm-based Fault Tolerance(ABFT) techniques can check the results of individual matrix multiplications. However, for a GCN layer, this check should be performed twice. To avoid this overhead, this work introduces GCN-ABFT that directly calculates a checksum for the entire three-matrix product within a single GCN layer, providing a cost-effective approach for error detection in GCN accelerators. Experimental results demonstrate that GCN-ABFT reduces the number of operations needed for checksum computation by over 21% on average for representative GCN applications. These savings are achieved without sacrificing fault-detection accuracy, as evidenced by the presented fault-injection analysis.

cs.AR↗

arXiv2024

Online Alignment and Addition in Multi-Term Floating-Point Adders

Multi-term floating-point addition appears in vector dot-product computations, matrix multiplications, and other forms of floating-point data aggregation. A critical step in multi-term floating point addition is the alignment of fractions of the floating-point terms before adding them. Alignment is executed serially by identifying first the maximum of all exponents and then shifting the fraction of each term according to the difference of its exponent from the maximum one. Contrary to common practice, this work proposes a new online algorithm that splits the identification of the maximum exponent, the alignment shift for each fraction, and their addition to multiple fused incremental steps that can be computed in parallel. Each fused step is implemented by a new associative operator that allows the incremental alignment and addition for arbitrary number of operands. Experimental results show that employing the proposed align-and-add operators for the implementation of multi-term floating point adders can improve delay or save significant area and power. The achieved area and power savings range between 3%-23% and 4%-26%, respectively.

cs.AR↗

arXiv2024

Floating-Point Multiply-Add with Approximate Normalization for Low-Cost Matrix Engines

The widespread adoption of machine learning algorithms necessitates hardware acceleration to ensure efficient performance. This acceleration relies on custom matrix engines that operate on full or reduced-precision floating-point arithmetic. However, conventional floating-point implementations can be power hungry. This paper proposes a method to improve the energy efficiency of the matrix engines used in machine learning algorithm acceleration. Our approach leverages approximate normalization within the floating-point multiply-add units as a means to reduce their hardware complexity, without sacrificing overall machine-learning model accuracy. Hardware synthesis results show that this technique reduces area and power consumption roughly by 16% and 13% on average for Bfloat16 format. Also, the error introduced in transformer model accuracy is 1% on average, for the most efficient configuration of the proposed approach.

cs.AR↗