FAME: An FPGA-Based Platform for Approximate Multipliers Evaluation with Pattern-Guided DNN Retraining

arXiv:2609.17730 · cs.AR, cs.LG, cs.PF · Submitted 2026-09-15 · Read on arXiv

cs.AR, cs.LG, cs.PF

Submitted: 2026-09-15

Updated: 2026-09-15

Comments: Accepted at the 38th IEEE/SBC International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD) 2026

Code: https://github.com/gicLAB/FAME

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

Accuracy — 106× in this paper · explained in 1 episode
Quantization — 29× in this paper · explained in 12 episodes
Core — 6× in this paper · explained in 2 episodes
Calibration — 5× in this paper · explained in 7 episodes
Deep Neural Network — 5× in this paper
Energy Efficiency — 4× in this paper · explained in 3 episodes
Partial Product — 4× in this paper
Benchmarking — 2× in this paper · explained in 4 episodes
Pattern Rounding Error — 2× in this paper
Quantization-Aware Training — 2× in this paper · explained in 3 episodes
Edge AI — 1× in this paper · explained in 2 episodes
Edge Devices — 1× in this paper · explained in 2 episodes
HLS Intellectual Property — 1× in this paper
Hardware Description Language — 1× in this paper
Power Delay Product — 1× in this paper
Robustness — 1× in this paper · explained in 5 episodes
Stochastic Gradient Descent (SGD) — 1× in this paper · explained in 5 episodes
Vivado High-Level Synthesis — 1× in this paper

Related papers