R.I.P.
๐ป
Ghosted
AsyncSparse: Accelerating Sparse Matrix-Matrix Multiplication on Asynchronous GPU Architectures
April 20, 2026 ยท Grace Period ยท + Add venue
Authors
Jie Liu, Huanzhi Pu, Zhiru Zhang
arXiv ID
2604.17834
Category
cs.DC: Distributed Computing
Citations
0
Abstract
Sparse Matrix-Matrix Multiplication (SpMM) is a fundamental kernel across scientific computing and machine learning. While prior work accelerates SpMM using Tensor Cores, no existing sparse kernel exploits the asynchronous features of modern GPU architectures, such as NVIDIA's Tensor Memory Accelerator (TMA) and warp specialization. This work systematically studies how these features impact SpMM performance and introduces two co-designed kernels. For structured sparsity, we optimize a warp-specialized producer-consumer pipeline overlapping TMA data transfer with WGMMA computation using Block Compressed Sparse Row (BCSR) format. For irregular sparsity, we design a Window Compressed Sparse Row (WCSR) kernel that loads the sparse operand via TMA and splits large row-windows across thread blocks for load balancing. Our WCSR kernel outperforms all prior SpMM kernels on SuiteSparse matrices (1.47x over AccSpMM, 6.24x over cuSPARSE). Our BCSR kernel achieves a combined 2.66x end-to-end speedup on Qwen2.5-7B prefill at 90% block sparsity with 64K tokens over cuDNN/cuBLAS.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
๐ Similar Papers
In the same crypt โ Distributed Computing
R.I.P.
๐ป
Ghosted
Reproducing GW150914: the first observation of gravitational waves from a binary black hole merger
R.I.P.
๐ป
Ghosted
MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
R.I.P.
๐ป
Ghosted
Adaptive Federated Learning in Resource Constrained Edge Computing Systems
R.I.P.
๐ป
Ghosted
Edge Intelligence: Paving the Last Mile of Artificial Intelligence with Edge Computing
R.I.P.
๐ป
Ghosted