# Performance Optimization of Deep Learning Sparse Matrix Kernels on Intel Max Series GPU

arXiv: https://arxiv.org/abs/2311.00368 · HTML: https://mlsanity.com/paper/2311.00368/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 19 references that are in the map.
- First posted on arXiv 1 Nov 2023.

## Builds on (most cited first, 6 of 6)

- [The State of Sparsity in Deep Neural Networks](https://mlsanity.com/paper/1902.09574/) (2019)
- [MegaBlocks: Efficient Sparse Training with Mixture-of-Experts](https://mlsanity.com/paper/2211.15841/) (2022)
- [Sparse GPU Kernels for Deep Learning](https://mlsanity.com/paper/2006.10901/) (2020)
- [FeatGraph: A Flexible and Efficient Backend for Graph Neural Network Systems](https://mlsanity.com/paper/2008.11359/) (2020)
- [FusedMM: A Unified SDDMM-SpMM Kernel for Graph Embedding and Graph Neural Networks](https://mlsanity.com/paper/2011.06391/) (2020)
- [Distributed-Memory Sparse Kernels for Machine Learning](https://mlsanity.com/paper/2203.07673/) (2022)

## Built on by (most cited first, 1 of 1)

- [Using Generative AI and Multi-Agents to Provide Automatic Feedback](https://mlsanity.com/paper/2411.07407/) (2024)

## Abstract (arXiv)

In this paper, we focus on three sparse matrix operations that are relevant for machine learning applications, namely, the sparse-dense matrix multiplication (SPMM), the sampled dense-dense matrix multiplication (SDDMM), and the composition of the SDDMM with SPMM, also termed as FusedMM. We develop optimized implementations for SPMM, SDDMM, and FusedMM operations utilizing Intel oneAPI's Explicit SIMD (ESIMD) SYCL extension API. In contrast to CUDA or SYCL, the ESIMD API enables the writing of explicitly vectorized kernel code. Sparse matrix algorithms implemented with the ESIMD API achieved performance close to the peak of the targeted Intel Data Center GPU. We compare our performance results to Intel's oneMKL library on Intel GPUs and to a recent CUDA implementation for the sparse matrix operations on NVIDIA's V100 GPU and demonstrate that our implementations for sparse matrix operations outperform either.