Skip to content

Latest commit

 

History

History
89 lines (59 loc) · 2.73 KB

README.md

File metadata and controls

89 lines (59 loc) · 2.73 KB
Raw

FGPUTensor

Copyright 2024 Fluid Numerics LLC

Build Status Build Status codecov

fgputensor is an equation parser Fortran class that is used to interpret and evaluate functions provided as strings. This library provides derived types to define "tensors" in Fortran that have CPU and GPU memory pointers.

CPU memory is defined through Fortran multi-dimensional pointer arrays GPU memory is defined through type(c_ptr) from iso_c_binding

Additional utilities are provided to make it a little easier to manage CPU and GPU data in Fortran.

Installation

fgputensor can be installed using CMake.

Prerequisites

All you need is a Fortran compiler that is compliant with the Fortran 2008 standard and supports C interoperability. Additionally, the table below lists the supported compilers. You will need to have CMake version 3.0.2 or greater.

CMake

For a quick installation to /usr/local/fgputensor,

cd build/
cmake ../
make
sudo make install

If you'd like to run the provided tests to verify your installation, simply run ctest while in the build/ subdirectory.

Supported Compilers

The following combinations are tested on the main branch of fgputensor:

Name Version Platform Build System Architecture
GNU Fortran 9, 10, 11, 12 Ubuntu 22.04.2 LTS cmake x86_64
Intel oneAPI (ifx) 2023.2 Ubuntu 22.04.2 LTS cmake x86_64
Intel oneAPI classic (ifort) 2021.1 Ubuntu 22.04.2 LTS cmake x86_64

Usage

Run examples

Note

Examples are included in the example/ subdirectory

Included examples

  • simple_2d_array.f90 - Creates a 2d float32 tensor with CPU and GPU data pointers.

Simple example with Makefile

Example Makefile

FC = gfortran
FLIBS += -L/usr/local/fgputensor/lib -lfgputensor
FFLAGS += -I/usr/local/fgputensor/include

demo : fgputensor_demo.f90
	${FC} -c fgputensor_demo.f90 ${FFLAGS}
	${FC} fgputensor_demo.o ${FFLAGS} ${FLIBS} -o $@

Example program

program fgputensor_demo

  use FGPUTensor

  implicit none

  type(tensor_f32_r2) :: t

  call t % init((/100,100/),0)

  call t % info()

  call t % free()

end program fgputensor_demo

Contributors

  • (Maintainer) Joe Schoonover, Fluid Numerics LLC