CsyrkEx

cuBLAS Extension APIs - cublasCsyrkEx

Description

This code demonstrates a usage of cuBLAS CSyrkEx function to perform a symmetric rank-k update

A = | 1.1 + 1.2j | 3.5 + 3.6j |
    | 3.5 + 3.6j | 4.7 + 4.8j |

This function is an extension of cublasCsyrk where the input matrix and output matrix can have a lower precision but the computation is still done in the type cuComplex.

See documentation for further details.

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

• cublasCsyrkEx API

Building (make)

Prerequisites

• A Linux/Windows system with recent NVIDIA drivers.
• CMake version 3.18 minimum

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 cublas_examples.sln project in Visual Studio and build

Usage

$  ./cublas_CsyrkEx_example

Sample example output:

A
1.10 + 1.20j 3.50 + 3.60j 
3.50 + 3.60j 4.70 + 4.80j 
=====
C
-28.78 + 26.90j -43.18 + 40.58j 
0.00 + 0.00j -71.98 + 68.66j 
=====