English
Related papers

Related papers: Flexible Performant GEMM Kernels on GPUs

200 papers

Emerging deep learning workloads urgently need fast general matrix multiplication (GEMM). To meet such demand, one of the critical features of machine-learning-specific accelerators such as NVIDIA Tensor Cores, AMD Matrix Cores, and Google…

Distributed, Parallel, and Cluster Computing · Computer Science 2023-11-13 Bo Fang , Xinyi Li , Harvey Dam , Cheng Tan , Siva Kumar Sastry Hari , Timothy Tsai , Ignacio Laguna , Dingwen Tao , Ganesh Gopalakrishnan , Prashant Nair , Kevin Barker , Ang Li

High performance dense linear algebra (DLA) libraries often rely on a general matrix multiply (Gemm) kernel that is implemented using assembly or with vector intrinsics. In particular, the real-valued Gemm kernels provide the overwhelming…

Mathematical Software · Computer Science 2017-05-01 Richard Michael Veras , Tze Meng Low , Tyler Michael Smith , Robert van de Geijn , Franz Franchetti

General Matrix Multiplication (GEMM) is a critical operation underpinning a wide range of applications in high-performance computing (HPC) and artificial intelligence (AI). The emergence of hardware optimized for low-precision arithmetic…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-08-21 Qiao Zhang , Rabab Alomairy , Dali Wang , Zhuowei Gu , Qinglei Cao

Developing kernels for Processing-In-Memory (PIM) platforms poses unique challenges in data management and parallel programming on limited processing units. Although software development kits (SDKs) for PIM, such as the UPMEM SDK, provide…

Hardware Architecture · Computer Science 2025-10-21 Krystian Chmielewski , Jarosław Ławnicki , Uladzislau Lukyanau , Tadeusz Kobus , Maciej Maciejewski

FPGAs are a promising platform for accelerating Deep Learning (DL) applications, due to their high performance, low power consumption, and reconfigurability. Recently, the leading FPGA vendors have enhanced their architectures to more…

Hardware Architecture · Computer Science 2024-04-18 Endri Taka , Dimitrios Gourounas , Andreas Gerstlauer , Diana Marculescu , Aman Arora

GPUs are popular devices for accelerating scientific calculations. However, as GPU code is usually written in low-level languages, it breaks the abstractions of high-level languages popular with scientific programmers. To overcome this, we…

Distributed, Parallel, and Cluster Computing · Computer Science 2016-04-13 Tim Besard , Pieter Verstraete , Bjorn De Sutter

Transformer training systems are built around dense linear algebra, yet a nontrivial fraction of end-to-end time is spent on surrounding memory-bound operators. Normalization, activations, residual updates, reductions, and related…

Machine Learning · Computer Science 2026-05-21 Han Guo , Jack Zhang , Arjun Menon , Driss Guessous , Vijay Thakkar , Yoon Kim , Tri Dao

In high-performance computing, hotspot GPU kernels are primary bottlenecks, and expert manual tuning is costly and hard to port. Large language model methods often assume kernels can be compiled and executed cheaply, which fails in large…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-12-30 Ruifan Chu , Anbang Wang , Xiuxiu Bai , Shuai Liu , Xiaoshe Dong

The KeOps library provides a fast and memory-efficient GPU support for tensors whose entries are given by a mathematical formula, such as kernel and distance matrices. KeOps alleviates the major bottleneck of tensor-centric libraries for…

Machine Learning · Computer Science 2021-04-10 Benjamin Charlier , Jean Feydy , Joan Alexis Glaunès , François-David Collin , Ghislain Durif

GEneral Matrix Multiplications (GEMMs) are recurrent in high-performance computing and deep learning workloads. Typically, high-end CPUs accelerate GEMM workloads with Single-Instruction Multiple Data (SIMD) or vector Instruction Set…

Hardware Architecture · Computer Science 2025-07-08 Alexandre de Limas Santana , Adrià Armejach , Francesc Martinez , Erich Focht , Marc Casas

Matrix libraries often focus on achieving high performance for problems considered to be either "small" or "large", as these two scenarios tend to respond best to different optimization strategies. We propose a unified technique for…

Mathematical Software · Computer Science 2023-02-20 RuQing G. Xu , Field G. Van Zee , Robert A. van de Geijn

Modelling of multivariate densities is a core component in many signal processing, pattern recognition and machine learning applications. The modelling is often done via Gaussian mixture models (GMMs), which use computationally expensive…

Mathematical Software · Computer Science 2018-05-10 Conrad Sanderson , Ryan Curtin

The generic matrix multiply (GEMM) function is the core element of high-performance linear algebra libraries used in many computationally-demanding digital signal processing (DSP) systems. We propose an acceleration technique for GEMM based…

Mathematical Software · Computer Science 2015-05-30 Davide Anastasia , Yiannis Andreopoulos

The devices designed for the Internet-of-Things encompass a large variety of distinct processor architectures, forming a highly heterogeneous zoo. In order to tackle this, we employ a simulator to estimate the performance of the…

Hardware Architecture · Computer Science 2024-03-13 Cristian Ramírez , Adrián Castelló , Héctor Martínez , Enrique S. Quintana-Ortí

Recent advances in self-supervised learning and the Transformer architecture have significantly improved natural language processing (NLP), achieving remarkably low perplexity. However, the growing size of NLP models introduces a memory…

Distributed, Parallel, and Cluster Computing · Computer Science 2024-04-02 Gunho Park , Baeseong Park , Minsub Kim , Sungjae Lee , Jeonghoon Kim , Beomseok Kwon , Se Jung Kwon , Byeongwook Kim , Youngjoo Lee , Dongsoo Lee

General matrix-matrix multiplication (GEMM) is a fundamental operation in machine learning (ML) applications. We present the first comprehensive performance acceleration of GEMM workloads on AMD's second-generation AIE-ML (AIE2)…

Hardware Architecture · Computer Science 2025-09-01 Kaustubh Mhatre , Endri Taka , Aman Arora

General Matrix Multiplication (GEMM) is a ubiquitous compute kernel in deep learning (DL). To support energy-efficient edge-native processing, new GEMM hardware units have been proposed that operate on unary encoded bitstreams using much…

Hardware Architecture · Computer Science 2024-12-25 Prabhu Vellaisamy , Harideep Nair , Joseph Finn , Manav Trivedi , Albert Chen , Anna Li , Tsung-Han Lin , Perry Wang , Shawn Blanton , John Paul Shen

General Matrix Multiplication (GEMM) is a fundamental operation in many scientific workloads, signal processing, and particularly deep learning. It is often a bottleneck for performance and energy efficiency, especially in edge environments…

Hardware Architecture · Computer Science 2025-11-11 Ilias Papalamprou , Dimosthenis Masouros , Ioannis Loudaros , Francky Catthoor , Dimitrios Soudris

Deep neural networks (DNNs) face significant challenges when deployed on resource-constrained extreme edge devices due to their computational and data-intensive nature. While standalone accelerators tailored for specific application…

Hardware Architecture · Computer Science 2024-11-22 Xiaoling Yi , Ryan Antonio , Joren Dumoulin , Jiacong Sun , Josse Van Delm , Guilherme Paim , Marian Verhelst

Modern accelerators like GPUs are increasingly executing independent operations concurrently to improve the device's compute utilization. However, effectively harnessing it on GPUs for important primitives such as general matrix…

Hardware Architecture · Computer Science 2024-09-05 Suchita Pati , Shaizeen Aga , Nuwan Jayasena , Matthew D. Sinclair