NIFA: Nonlinear IMC enhanced FPGA for efficient ML inference

• URL: http://arxiv.org/abs/2607.15123v1
• Date: Thu, 16 Jul 2026 15:32:44 GMT
• Title: NIFA: Nonlinear IMC enhanced FPGA for efficient ML inference
• Abstract summary: We propose a novel FPGA architecture that integrates an ADC-free IMC block, replacing the conventional ADC with analog content-addressable memories (ACAMs)<n>On CNN and Transformer-based benchmarks, the proposed architecture achieves up to 40x and 1.9x higher energy efficiency and 4.1x and 2.5x higher area efficiency, respectively.
• Score: 4.620721969034322
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Recent FPGAs have improved deep learning (DL) inference efficiency through dedicated tensor blocks and in-BRAM computation. ReRAM-based analog in-memory computing (IMC) pushes efficiency further, offering an order-of-magnitude improvement in compute density and energy efficiency over conventional digital logic by performing vector-matrix multiplication (VMM) directly within the ReRAM crossbar; prior work has integrated such IMC blocks into FPGAs for DL inference. However, conventional IMC designs support only static-weight VMM, leaving nonlinear operations and dynamic matrix-matrix multiplication (DIMM) to the FPGA fabric. As a result, the benefits of IMC are largely confined to static-weight models, whereas Transformer-based models, which rely on frequent nonlinear and DIMM operations, gain only limited improvement. Moreover, the ADCs within each IMC block consume more than 70% of its area and power, further limiting system efficiency and scalability. To address these limitations, we propose a novel FPGA architecture that integrates an ADC-free IMC block, replacing the conventional ADC with analog content-addressable memories (ACAMs) that natively perform nonlinear operations inside the block. To fully exploit this block, we conduct an FPGA-aware design-space exploration that determines optimal crossbar dimensions while balancing FPGA area, flexibility, and DL performance, and we develop an efficient mapping that leverages ACAMs to carry out DIMM operations, extending the applicability of IMC to attention computation. On CNN and Transformer-based benchmarks, the proposed architecture achieves up to 40x and 1.9x higher energy efficiency and 4.1x and 2.5x higher area efficiency, respectively. Overall, it significantly improves FPGA DL inference efficiency and sustains robust gains on Transformer-based workloads across long input sequences, advancing domain-specialized FPGA design.

Related papers

• High-Performance NTT Accelerators for PQC leveraging Unified Redundant Arithmetic and Fine-Tuned Microarchitecture [0.3186130813218338]
  Post-quantum cryptography and privacy-preserving technologies are expected to play a central role in future secure communication systems.<n>We introduce a novel number representation that eliminates conditional corrections for both Montgomery modulo multiplication and combined subtract-multiply operations.
  arXiv  Detail & Related papers  (2026-07-01T08:41:11Z)
• CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs [48.961770224590204]
  CODA is a GPU kernel abstraction that expresses these computations as GEMM-plus-epilogue programs.<n>We show that CODA offers a practical path toward combining framework-level productivity with hardware-level efficiency.
  arXiv  Detail & Related papers  (2026-05-19T02:30:43Z)
• AdaFuse: Accelerating Dynamic Adapter Inference via Token-Level Pre-Gating and Fused Kernel Optimization [84.25316984309725]
  Integration of dynamic, sparse structures with parameter-efficient adapters (e.g., LoRA) is a powerful technique for enhancing Large Language Models (LLMs)<n>Despite minimal increases in computational load, the computation latency often skyrockets, leading to decoding speeds slowing by over 2.5 times.<n>AdaFuse is a framework built on a tight co-design between the algorithm and the underlying hardware system to enable efficient dynamic adapter execution.
  arXiv  Detail & Related papers  (2026-03-12T12:46:42Z)
• POET-X: Memory-efficient LLM Training by Scaling Orthogonal Transformation [57.57816409869894]
  We introduce POET-X, a scalable and memory-efficient variant for training large language models.<n>PoET-X maintains the generalization and stability benefits of POET while achieving substantial improvements in throughput and memory efficiency.
  arXiv  Detail & Related papers  (2026-03-05T18:59:23Z)
• Eliminating Multi-GPU Performance Taxes: A Systems Approach to Efficient Distributed LLMs [61.953548065938385]
  We introduce the ''Three Taxes'' (Bulk Synchronous, Inter- Kernel Data Locality, and Kernel Launch Overhead) as an analytical framework.<n>We propose moving beyond the rigid BSP model to address key inefficiencies in distributed GPU execution.<n>We observe a 10-20% speedup in end-to-end latency over BSP-based approaches.
  arXiv  Detail & Related papers  (2025-11-04T01:15:44Z)
• Joint Channel Estimation and Computation Offloading in Fluid Antenna-assisted MEC Networks [81.36647816787713]
  We propose an FA-assisted offloading framework to minimize the delay of channel estimation.<n>We show that the proposed system significantly reduces the accuracy under efficient communication.
  arXiv  Detail & Related papers  (2025-09-16T08:48:44Z)
• Design and Implementation of an FPGA-Based Hardware Accelerator for Transformer [0.0]
  Transformer-based large language models rely heavily on matrix multiplications for attention and feed-forward layers.<n>We introduce a highly optimized tiled matrix multiplication accelerator on a resource-constrained Xilinx KV260 FPGA.<n>Our design exploits persistent on-chip storage, a robust two-level tiling strategy for maximal data reuse, and a systolic-like unrolled compute engine.
  arXiv  Detail & Related papers  (2025-03-20T22:15:42Z)
• Runtime Tunable Tsetlin Machines for Edge Inference on eFPGAs [0.2294388534633318]
  eFPGAs allow for the design of hardware accelerators of edge Machine Learning (ML) applications at a lower power budget.<n>The limited eFPGA logic and memory significantly constrain compute capabilities and model size.<n>The proposed eFPGA accelerator focuses on minimizing resource usage and allowing flexibility for on-field recalibration over throughput.
  arXiv  Detail & Related papers  (2025-02-10T12:49:22Z)
• Fast, Scalable, Energy-Efficient Non-element-wise Matrix Multiplication on FPGA [10.630802853096462]
  Modern Neural Network (NN) architectures heavily rely on vast numbers of multiply-accumulate arithmetic operations. This paper proposes a high- throughput, scalable and energy efficient non-element-wise matrix multiplication unit on FPGAs. Using our AMU achieves up to 9x higher throughput and 112x higher energy efficiency over the state-of-the-art solutions for the FPGA-based Quantised Neural Network (QNN) accelerators.
  arXiv  Detail & Related papers  (2024-07-02T15:28:10Z)
• SWAT: Scalable and Efficient Window Attention-based Transformers Acceleration on FPGAs [3.302913401404089]
  Sliding window-based static sparse attention mitigates the problem by limiting the attention scope of the input tokens. We propose a dataflow-aware FPGA-based accelerator design, SWAT, that efficiently leverages the sparsity to achieve scalable performance for long input.
  arXiv  Detail & Related papers  (2024-05-27T10:25:08Z)
• Adaptable Butterfly Accelerator for Attention-based NNs via Hardware and Algorithm Co-design [66.39546326221176]
  Attention-based neural networks have become pervasive in many AI tasks. The use of the attention mechanism and feed-forward network (FFN) demands excessive computational and memory resources. This paper proposes a hardware-friendly variant that adopts a unified butterfly sparsity pattern to approximate both the attention mechanism and the FFNs.
  arXiv  Detail & Related papers  (2022-09-20T09:28:26Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.