High-Performance I/O for Scientific Applications
Robert Latham
Mathematics and Computer Science Division, 
Argonne National Lab, IL, USA


Abstract: 

For all their hype, parallel file systems alone are not sufficient for
achieving high-performance and usable I/O on large clusters.  In addition
to the performance that a parallel file system can afford, functionality
is also needed for mapping high-level or domain specific application
abstractions into file system ones and for managing the many processes
that make up these applications.

Because of these additional demands, high-performance I/O (HPIO) systems
in parallel machines have evolved into multi-tiered solutions.  At the top
are high-level I/O libraries such as HDF5 and PnetCDF that provide
application developers familiar interfaces and abstractions such as
multidimensional typed variables. Below this layer sits MPI-IO, providing
management of groups of processes and aggregation of operations performed
by these groups.  The parallel file system sits at the lowest software
layer and provides a unified view of a large number of I/O resources and
manages concurrent access to these resources.

This talk will focus on the application programmer's point of view.  We
will discuss how the layers of this software stack lead to an effective
solution, what happens to I/O requests as they move through this software
stack, the tradeoffs of interfacing at the various layers, and general
guidelines for obtaining high performance.  Attendees will come out of the
tutorial with an understanding of how HPIO systems are constructed and how
they are different from other storage systems, knowledge of what happens
in a HPIO system after an application makes an I/O call, a familiarity
with the tools available to them for making use of these resources, and
some guidelines for extracting performance from these complicated systems.

Table of Contents:
- Introduction and I/O stacks
    - Application I/O vs. parallel I/O 
    - Bridging the gap with I/O stacks 
    - I/O stacks for computational science 
        1.High Level Libraries
        2.I/O Middleware
        3.Parallel File System
- I/O interfaces and formats, with examples 
    - POSIX file system interface 
        1.Examples
        2.POSIX and PFS interaction
    - MPI-IO interface 
        1.Collective I/O
        2.Noncontiguous I/O
        3.Nonblocking and Asynchronous I/O
        4.Examples
        5.Optimizations
        6.MPI-IO Implementations
    - Parallel netCDF (PnetCDF) 
        1.File Layout
        2.Usage
        3.Examples
        4.Interaction with MPI-IO
    - Hierarchical Data Format (HDF5) 
        1.File Layout
        2.Usage
        3.Examples
        4.Interaction with MPI-IO
- I/O best practices 
    - Choosing an I/O interface 
    - Guidelines for I/O performance 
    - Tuning I/O stacks with hints 
        1.MPI-IO hints
        2.File system-specific hints
        3.High-level library hints
    - Enlisting the experts 
- Conclusions and supplemental material