# High-Performance Deep Learning via a Single Building Block

arXiv: https://arxiv.org/abs/1906.06440 · HTML: https://mlsanity.com/paper/1906.06440/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 15 Jun 2019.

## Built on by (most cited first, 2 of 2)

- [SparseDNN: Fast Sparse Deep Learning Inference on CPUs](https://mlsanity.com/paper/2101.07948/) (2021)
- [Efficient and Generic 1D Dilated Convolution Layer for Deep Learning](https://mlsanity.com/paper/2104.08002/) (2021)

## Abstract (arXiv)

Deep learning (DL) is one of the most prominent branches of machine learning. Due to the immense computational cost of DL workloads, industry and academia have developed DL libraries with highly-specialized kernels for each workload/architecture, leading to numerous, complex code-bases that strive for performance, yet they are hard to maintain and do not generalize. In this work, we introduce the batch-reduce GEMM kernel and show how the most popular DL algorithms can be formulated with this kernel as the basic building-block. Consequently, the DL library-development degenerates to mere (potentially automatic) tuning of loops around this sole optimized kernel. By exploiting our new kernel we implement Recurrent Neural Networks, Convolution Neural Networks and Multilayer Perceptron training and inference primitives in just 3K lines of high-level code. Our primitives outperform vendor-optimized libraries on multi-node CPU clusters, and we also provide proof-of-concept CNN kernels targeting GPUs. Finally, we demonstrate that the batch-reduce GEMM kernel within a tensor compiler yields high-performance CNN primitives, further amplifying the viability of our approach.