{"ID":2844268,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.11640","arxiv_id":"2511.11640","title":"Exploring Parallelism in FPGA-Based Accelerators for Machine Learning Applications","abstract":"Speculative backpropagation has emerged as a promising technique to accelerate the training of neural networks by overlapping the forward and backward passes. Leveraging speculative weight updates when error gradients fall within a specific threshold reduces training time without substantially compromising accuracy. In this work, we implement speculative backpropagation on the MNIST dataset using OpenMP as the parallel programming platform. OpenMP's multi-threading capabilities enable simultaneous execution of forward and speculative backpropagation steps, significantly improving training speed. The application is planned for synthesis on a state-of-the-art FPGA to demonstrate its potential for hardware acceleration. Our CPU-based experimental results demonstrate that speculative backpropagation achieves a maximum speedup of 24% in execution time when using a threshold of 0.25, and accuracy remaining within 3-4% of the baseline across various epochs. Additionally, when comparing individual step execution time, speculative backpropagation yields a maximum speedup of 35% over the baseline, demonstrating the effectiveness of overlapping forward and backward passes.","short_abstract":"Speculative backpropagation has emerged as a promising technique to accelerate the training of neural networks by overlapping the forward and backward passes. Leveraging speculative weight updates when error gradients fall within a specific threshold reduces training time without substantially compromising accuracy. In...","url_abs":"https://arxiv.org/abs/2511.11640","url_pdf":"https://arxiv.org/pdf/2511.11640v1","authors":"[\"Sed Centeno\",\"Christopher Sprague\",\"Arnab A Purkayastha\",\"Ray Simar\",\"Neeraj Magotra\"]","published":"2025-11-09T05:05:05Z","proceeding":"cs.DC","tasks":"[\"cs.DC\",\"cs.AR\",\"cs.LG\"]","methods":"[]","has_code":false}