Skip to content

Latest commit

 

History

History
17 lines (13 loc) · 1.14 KB

Readme.md

File metadata and controls

17 lines (13 loc) · 1.14 KB
Raw

CUDA Fortran 90 Feature Tests

Purpose

The purpose of this repository is to experiment with the basics of CUDA programming in Fortran 90. This repository consists of several small projects to test speed ups, memory allocation, streaming, etc. It can be used as a basic refernce for using CUDA programming in modern Fortran. It is noteworthy to mention that some of the examples are taken directly from the [PGI CUDA Fortran Programming Guide][https://www.pgroup.com/doc/pgi17cudaforug.pdf]; these folders carry the example in their directory name.

Contents

  • 01-get-threadid
  • example-5.1: An extensive example for large matrix by matrix multiplication using CPU (double loop or OpenBLAS) and GPU (cuBLAS and slicing).
  • example-5.2: replica of the mapped-memory allocation
  • memory-bandwidth: measures the effective Host2Device and Device2Host transfers for the pinned versus pageable memory.
  • async-data-transfer: measures the latency of four different data transfer strategies

Requirements

References