18943

High-Performance Deep Learning via a Single Building Block

Evangelos Georganas, Kunal Banerjee, Dhiraj Kalamkar, Sasikanth Avancha, Anand Venkat, Michael Anderson, Greg Henry, Hans Pabst, Alexander Heinecke
Intel Corporation
arXiv:1906.06440 [cs.LG], (18 Jun 2019)

@misc{georganas2019highperformance,

   title={High-Performance Deep Learning via a Single Building Block},

   author={Evangelos Georganas and Kunal Banerjee and Dhiraj Kalamkar and Sasikanth Avancha and Anand Venkat and Michael Anderson and Greg Henry and Hans Pabst and Alexander Heinecke},

   year={2019},

   eprint={1906.06440},

   archivePrefix={arXiv},

   primaryClass={cs.LG}

}

Download Download (PDF)   View View   Source Source   

401

views

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.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2019 hgpu.org

All rights belong to the respective authors

Contact us: