Stream-K: Work-centric Parallel Decomposition for Dense Matrix-Matrix Multiplication on the GPU
January 09, 2023 Β· Declared Dead Β· π ACM SIGPLAN Symposium on Principles & Practice of Parallel Programming
"No code URL or promise found in abstract"
Evidence collected by the PWNC Scanner
Authors
Muhammad Osama, Duane Merrill, Cris Cecka, Michael Garland, John D. Owens
arXiv ID
2301.03598
Category
cs.DS: Data Structures & Algorithms
Cross-listed
cs.DC
Citations
37
Venue
ACM SIGPLAN Symposium on Principles & Practice of Parallel Programming
Last Checked
3 months ago
Abstract
We introduce Stream-K, a work-centric parallelization of matrix multiplication (GEMM) and related computations in dense linear algebra. Whereas contemporary decompositions are primarily tile-based, our method operates by partitioning an even share of the aggregate inner loop iterations among physical processing elements. This provides a near-perfect utilization of computing resources, regardless of how efficiently the output tiling for any given problem quantizes across the underlying processing elements. On GPU processors, our Stream-K parallelization of GEMM produces a peak speedup of up to 14$\times$ and 6.7$\times$, and an average performance response that is both higher and more consistent across 32,824 GEMM problem geometries than state-of-the-art math libraries such as CUTLASS and cuBLAS. Furthermore, we achieve this performance from a single tile size configuration per floating-point precision, whereas today's math libraries employ complex kernel-selection heuristics to select from a large ensemble of kernel variants.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
π Similar Papers
In the same crypt β Data Structures & Algorithms
π
π
The Cartographer
R.I.P.
π»
Ghosted
Route Planning in Transportation Networks
R.I.P.
π»
Ghosted
Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration
R.I.P.
π»
Ghosted
Hierarchical Clustering: Objective Functions and Algorithms
R.I.P.
π»
Ghosted
Graph Isomorphism in Quasipolynomial Time
π
π
The Cartographer
Simulation optimization: A review of algorithms and applications
Died the same way β π» Ghosted
R.I.P.
π»
Ghosted
Federated Learning: Strategies for Improving Communication Efficiency
R.I.P.
π»
Ghosted
In-Datacenter Performance Analysis of a Tensor Processing Unit
R.I.P.
π»
Ghosted
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning
R.I.P.
π»
Ghosted