Enabling Memory-efficient Im2win Convolution with Multi-precision Support on GPU CUDA and Tensor Cores

August 21, 2026 ยท Grace Period ยท ๐Ÿ› the 2026 International Joint Conference on Neural Networks

โณ Grace Period
This paper is less than 90 days old. We give authors time to release their code before passing judgment.
Authors Xiang Fu, Jixiang Ma, Xinpeng Zhang, Peng Zhao, Shuai Lu, Xu Tony Liu arXiv ID 2608.20725 Category cs.DC: Distributed Computing Cross-listed cs.CV Citations 0 Venue the 2026 International Joint Conference on Neural Networks
Abstract
Convolution is a principal computational bottleneck in deep neural networks, and its efficiency depends on tight integration between algorithms and GPU hardware. Existing GPU convolution methods suffer from large memory overhead, poor cache utilization, limited effectiveness across kernel sizes, or numerical instability. This work extends the im2win paradigm -- a universal, memory-efficient convolution method with contiguous memory access for all kernel sizes -- to run efficiently in full precision on CUDA cores and half precision on tensor cores. By introducing new kernel designs and optimizations such as zig-zag memory access and asynchronous data movement, im2win efficiently exploits hardware-accelerated half-precision matrix multiply-accumulate operations. Across twelve CNN benchmarks, im2win achieves up to 2.8x higher TFLOPS than its CUDA core implementation, 1.4x higher than cuDNN, and 6.4x higher than GEMM-based convolution with cuBLAS, while using as little as 53% and 35% of their memory, respectively. These results establish im2win as a unified, high-performance convolution framework for modern GPU architectures.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Distributed Computing