GemmEx

cuBLAS Extension APIs - cublasGemmEx

Description

This code demonstrates a usage of cuBLAS GemmEx function to matrix-matrix product

A = | 1.0 | 2.0 |
    | 3.0 | 4.0 |

B = | 5.0 | 6.0 |
    | 7.0 | 8.0 |

This function is an extension of cublas<t>gemm that allows the user to individually specify the data types for each of the A, B and C matrices, the precision of computation and the GEMM algorithm to be run. Supported combinations of arguments are listed further down in this section.

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

• cublasGemmEx 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_GemmEx_example

Sample example output:

A
1.00 2.00 
3.00 4.00 
=====
B
5.00 6.00 
7.00 8.00 
=====
C
19.00 22.00 
43.00 50.00 
=====