burlen / h3d-io Goto Github PK

!      ____    _ __           ____               __    ____
!     / __/___(_) /  ___ ____/ __ \__ _____ ___ / /_  /  _/__  ____
!    _\ \/ __/ / _ \/ -_) __/ /_/ / // / -_|_-</ __/ _/ // _ \/ __/
!   /___/\__/_/_.__/\__/_/  \___\_\_,_/\__/___/\__/ /___/_//_/\__(_) 
!
!
! Copyright 2010 SciberQuest Inc.
!
! No permission is granted to reproduce this software.
!
! This is experimental software and is provided ‘‘as is’’, with no
! warranties of any kind whatsoever, no support, no promise of updates,
! or printed documentation.
!==============================================================================

About H3D-IO
----------------------------
A parallel IO library optimized for the Lustre filesystem providing IO on
cartesian grids for analysis data and IO in contiguous chunk pattern for 
checkpoint dumps. The library's was develoed to meet the needs of the 
UC San Diego H3D global hybrid magnetospheric simulation code. The library
comes with two benchmark codes that were used to fine tune its performance
during development.

Building
----------------------------
# configure [system-name]
# make

configure with no options lists available system-names. 
On Kraken use "kraken-lustre".
On a linux system without a Lustre filesystem use "linux-mpich" or "linux-mpich-debug". 

Running Benchmarks 
----------------------------
# ./IOBenchmarkContiguous /path/to/run/config /path/to/log/file /path/to/scratch/dir 
# ./IOBenchmarkUniformGrid /path/to/run/config /path/to/log/file /path/to/scratch/dir

Notes
----------------------------
Has been tested using gcc and Intel compilers. It compiled with PGI compilers
at one point, however it hasn't been run extensively with PGI compilers.