Search papers, labs, and topics across Lattice.
MINT is a dynamic-precision CNN inference accelerator that employs most-significant-digit-first (MSDF) arithmetic to optimize computation by exposing partial results early, allowing for early termination once the desired precision is reached. By conducting a budget-constrained greedy search, MINT selects the lowest precision for each convolution layer while maintaining accuracy within 2% of the INT8 baseline for VGG-16 and ResNet-18 networks. The implementation on a Xilinx Zynq-7020 achieves significantly higher throughput and energy efficiency compared to traditional INT8 methods, with minimal accuracy loss.
Achieving up to 82% higher energy efficiency while maintaining accuracy, MINT redefines the boundaries of CNN inference on FPGA platforms.
We present MINT, a dynamic-precision CNN inference accelerator based on left-to-right (LR) arithmetic. LR arithmetic computes in most-significant-digit-first manner and exposes useful partial results early so that the computation can be terminated once the desired precision is achieved. At the core, there is a MSDF serial-parallel inner-product unit, which uses redundant signed-digit representation to compute each convolution window. A budget-constrained greedy search profiles all convolution layers from INT2 to INT7 and selects the lowest precision per layer while constraining total accuracy loss to within 2\% of the INT8 baseline for VGG-16 and ResNet-18 networks. The design is synthesized on a Xilinx Zynq-7020 at \SI{200}{\mega\hertz}, and uses 5.64 average bits for VGG-16 and 6.04 for ResNet-18, while achieving 19.86 GOPS and 29.51 GOPS/W on VGG-16, and 18.86 GOPS and 26.40 GOPS/W on ResNet-18. This corresponds to 32.6\% and 26.0\% higher throughput and 82.10\% and 62.90\% higher energy efficiency than INT8 with only 1.81\% and 1.96\% drops relative to the INT8 baseline. Compared with representative prior FPGA CNN accelerators considered in this study, MINT delivers the highest energy efficiency among the listed VGG-16 and ResNet-18 designs on Zynq-7020 platform.
