A Parallel Scan Algorithm in the Tensor Core Unit Model

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zouzias, Anastasios, McColl, William F.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909406085513216
author Zouzias, Anastasios
McColl, William F.
author_facet Zouzias, Anastasios
McColl, William F.
contents We present a parallel scan (prefix sum) algorithm in the Tensor Core Unit (TCU) model of computation. The TCU model assumes that multiplication between two square matrices of constant size $s$ is a basic operation. In the $(s^2, \ell)$-TCU model, we show that for inputs of size $n$, the algorithm has depth at most $2\lfloor \log_s (n)\rfloor$ and runs in $O(n(1 + \ell /s^2)/p + (s^2 + \ell) \log_s (n))$ time assuming $p$ tensor core units. Equivalently, the algorithm performs $O(n/s^2)$ multiplications of square matrices of size s.
format Preprint
id arxiv_https___arxiv_org_abs_2411_17887
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Parallel Scan Algorithm in the Tensor Core Unit Model
Zouzias, Anastasios
McColl, William F.
Distributed, Parallel, and Cluster Computing
Data Structures and Algorithms
We present a parallel scan (prefix sum) algorithm in the Tensor Core Unit (TCU) model of computation. The TCU model assumes that multiplication between two square matrices of constant size $s$ is a basic operation. In the $(s^2, \ell)$-TCU model, we show that for inputs of size $n$, the algorithm has depth at most $2\lfloor \log_s (n)\rfloor$ and runs in $O(n(1 + \ell /s^2)/p + (s^2 + \ell) \log_s (n))$ time assuming $p$ tensor core units. Equivalently, the algorithm performs $O(n/s^2)$ multiplications of square matrices of size s.
title A Parallel Scan Algorithm in the Tensor Core Unit Model
topic Distributed, Parallel, and Cluster Computing
Data Structures and Algorithms
url https://arxiv.org/abs/2411.17887