gesvdaStridedBatched

cuSOLVER Approximate Singular Value Decomposition example

Description

This code demonstrates a usage of cuSOLVER gesvdaStridedBatched function to approximate singular value decomposition by gesvdaStridedBatched.

A = U * Σ * V^H

A0 and A1 are a 3x2 dense matrices,

A0 = | 1.0 | 2.0 |
     | 4.0 | 5.0 |
     | 2.0 | 1.0 |

A1 = | 10.0 | 9.0 |
     |  8.0 | 7.0 |
     |  6.0 | 5.0 |

Supported SM Architectures

All GPUs supported by CUDA Toolkit (https://developer.nvidia.com/cuda-gpus)

Supported OSes

Linux
Windows

Supported CPU Architecture

x86_64
ppc64le
arm64-sbsa

CUDA APIs involved

• cusolverDnSgesvdaStridedBatched_bufferSize API
• cusolverDnSgesvdaStridedBatched API

Building (make)

Prerequisites

• A Linux/Windows system with recent NVIDIA drivers.
• CMake version 3.18 minimum
• Minimum CUDA 10.1 toolkit is required.

Build command on Linux

$ mkdir build
$ cd build
$ cmake ..
$ make

Make sure that CMake finds expected CUDA Toolkit. If that is not the case you can add argument -DCMAKE_CUDA_COMPILER=/path/to/cuda/bin/nvcc to cmake command.

Build command on Windows

$ mkdir build
$ cd build
$ cmake -DCMAKE_GENERATOR_PLATFORM=x64 ..
$ Open cusolver_examples.sln project in Visual Studio and build

Usage

$  ./cusolver_gesvdaStridedBatched_example

Sample example output:

A0 = (matlab base-1)
1.00 2.00
4.00 5.00
2.00 1.00
=====
A1 = (matlab base-1)
10.00 9.00
8.00 7.00
6.00 5.00
=====
0-th matrix, gesvda converges
1-th matrix, gesvda converges
S0 = (matlab base-1)
7.07
1.04
=====
U0 = (matlab base-1)
0.31 -0.49
0.91 -0.11
0.29 0.87
=====
V) = (matlab base-1)
0.64 0.77
0.77 -0.64
=====
|S0 - S0_exact|_sup = 0.000000E+00
residual |A0 - U0*S0*V0**H|_F = 4.768372E-07
S1 = (matlab base-1)
18.84
0.26
=====
U1 = (matlab base-1)
0.71 0.57
0.56 -0.12
0.41 -0.81
=====
V1 = (matlab base-1)
0.75 -0.66
0.66 0.75
=====
|S1 - S1_exact|_sup = 0.000000E+00
residual |A1 - U1*S1*V1**H|_F = 0.000000E+00