代码拉取完成,页面将自动刷新
cff-version: 1.2.0
title: CUTLASS
message: >-
If you use this software, please cite using the
following metadata.
type: software
authors:
- given-names: Vijay
family-names: Thakkar
email: vithakkar@nvidia.com
affiliation: NVIDIA
- given-names: Pradeep
family-names: Ramani
email: prramani@nvidia.com
affiliation: NVIDIA
- given-names: Cris
family-names: Cecka
email: ccecka@nvidia.com
affiliation: NVIDIA
- given-names: Aniket
family-names: Shivam
email: ashivam@nvidia.com
affiliation: NVIDIA
- given-names: Honghao
family-names: Lu
email: honghaol@nvidia.com
affiliation: NVIDIA
- given-names: Ethan
family-names: Yan
email: etyan@nvidia.com
affiliation: NVIDIA
- given-names: Jack
family-names: Kosaian
email: jkosaian@nvidia.com
affiliation: NVIDIA
- given-names: Mark
family-names: Hoemmen
email: mhoemmen@nvidia.com
affiliation: NVIDIA
- given-names: Haicheng
family-names: Wu
email: haichengw@nvidia.com
affiliation: NVIDIA
- given-names: Andrew
family-names: Kerr
email: akerr@nvidia.com
affiliation: NVIDIA
- given-names: Matt
family-names: Nicely
email: mnicely@nvidia.com
affiliation: NVIDIA
- given-names: Duane
family-names: Merrill
email: dumerrill@nvidia.com
affiliation: NVIDIA
- given-names: Dustyn
family-names: Blasig
email: dblasig@nvidia.com
affiliation: NVIDIA
- given-names: Fengqi
family-names: Qiao
email: fqiao@nvidia.com
affiliation: NVIDIA
- given-names: Piotr
family-names: Majcher
email: pmajcher@nvidia.com
affiliation: NVIDIA
- given-names: Paul
family-names: Springer
email: pspringer@nvidia.com
affiliation: NVIDIA
- given-names: Markus
family-names: Hohnerbach
affiliation: NVIDIA
email: mhohnerbach@nvidia.com
- given-names: Jin
family-names: Wang
email: jinw@nvidia.com
affiliation: NVIDIA
- given-names: Manish
family-names: Gupta
affiliation: Google
email: manigupta@google.com
repository-code: 'https://github.com/NVIDIA/cutlass'
abstract: >-
CUTLASS is a collection of CUDA C++ template
abstractions for implementing high-performance
matrix-multiplication (GEMM) and related
computations at all levels and scales within CUDA.
It incorporates strategies for hierarchical
decomposition and data movement similar to those
used to implement cuBLAS and cuDNN. CUTLASS
decomposes these "moving parts" into reusable,
modular software components abstracted by C++
template classes. These thread-wide, warp-wide,
block-wide, and device-wide primitives can be
specialized and tuned via custom tiling sizes, data
types, and other algorithmic policy. The resulting
flexibility simplifies their use as building blocks
within custom kernels and applications.
keywords:
- 'cutlass, tensor cores, cuda, cute, nvidia, gpu, linear algebra, matrix computations'
license: BSD-3-Clause
license-url: https://github.com/NVIDIA/cutlass/blob/v3.0.0/LICENSE.txt
version: '3.0.0'
date-released: '2023-01-23'
identifiers:
- type: url
value: "https://github.com/NVIDIA/cutlass/tree/v3.0.0"
description: The GitHub release URL of tag 3.0.0
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。