starpu-max/doc/doxygen/chapters/000_introduction.doxy

/*
 * This file is part of the StarPU Handbook.
 * Copyright (C) 2009--2011  Universit@'e de Bordeaux
 * Copyright (C) 2010, 2011, 2012, 2013, 2014, 2016  CNRS
 * Copyright (C) 2011, 2012, 2016 INRIA
 * See the file version.doxy for copying conditions.
*/

/*! \mainpage Introduction

\htmlonly
<h1><a class="anchor" id="Foreword"></a>Foreword</h1>
\endhtmlonly
\htmlinclude version.html
\htmlinclude foreword.html

\section Motivation Motivation

// This is a comment and it will be removed before the file is processed by doxygen
// complex machines with heterogeneous cores/devices

The use of specialized hardware such as accelerators or coprocessors offers an
interesting approach to overcome the physical limits encountered by processor
architects. As a result, many machines are now equipped with one or several
accelerators (e.g. a GPU), in addition to the usual processor(s). While a lot of
efforts have been devoted to offload computation onto such accelerators, very
little attention as been paid to portability concerns on the one hand, and to the
possibility of having heterogeneous accelerators and processors to interact on the other hand.

StarPU is a runtime system that offers support for heterogeneous multicore
architectures, it not only offers a unified view of the computational resources
(i.e. CPUs and accelerators at the same time), but it also takes care of
efficiently mapping and executing tasks onto an heterogeneous machine while
transparently handling low-level issues such as data transfers in a portable
fashion.

// this leads to a complicated distributed memory design
// which is not (easily) manageable by hand
// added value/benefits of StarPU
//    - portability
//   - scheduling, perf. portability

\section StarPUInANutshell StarPU in a Nutshell

StarPU is a software tool aiming to allow programmers to exploit the
computing power of the available CPUs and GPUs, while relieving them
from the need to specially adapt their programs to the target machine
and processing units.

At the core of StarPU is its run-time support library, which is
responsible for scheduling application-provided tasks on heterogeneous
CPU/GPU machines.  In addition, StarPU comes with programming language
support, in the form of extensions to languages of the C family
(\ref cExtensions), as well as an OpenCL front-end (\ref SOCLOpenclExtensions).

StarPU's run-time and programming language extensions support a
task-based programming model. Applications submit computational
tasks, with CPU and/or GPU implementations, and StarPU schedules these
tasks and associated data transfers on available CPUs and GPUs.  The
data that a task manipulates are automatically transferred among
accelerators and the main memory, so that programmers are freed from the
scheduling issues and technical details associated with these transfers.

StarPU takes particular care of scheduling tasks efficiently, using
well-known algorithms from the literature (\ref TaskSchedulingPolicy).
In addition, it allows scheduling experts, such as compiler or
computational library developers, to implement custom scheduling
policies in a portable fashion (\ref DefiningANewSchedulingPolicy).

The remainder of this section describes the main concepts used in StarPU.

// explain the notion of codelet and task (i.e. g(A, B)

\subsection CodeletAndTasks Codelet and Tasks

One of the StarPU primary data structures is the \b codelet. A codelet describes a
computational kernel that can possibly be implemented on multiple architectures
such as a CPU, a CUDA device or an OpenCL device.

// TODO insert illustration f: f_spu, f_cpu, ...

Another important data structure is the \b task. Executing a StarPU task
consists in applying a codelet on a data set, on one of the architectures on
which the codelet is implemented. A task thus describes the codelet that it
uses, but also which data are accessed, and how they are
accessed during the computation (read and/or write).
StarPU tasks are asynchronous: submitting a task to StarPU is a non-blocking
operation. The task structure can also specify a \b callback function that is
called once StarPU has properly executed the task. It also contains optional
fields that the application may use to give hints to the scheduler (such as
priority levels).

By default, task dependencies are inferred from data dependency (sequential
coherency) by StarPU. The application can however disable sequential coherency
for some data, and dependencies can be specifically expressed.
A task may be identified by a unique 64-bit number chosen by the application
which we refer as a \b tag.
Task dependencies can be enforced either by the means of callback functions, by
submitting other tasks, or by expressing dependencies
between tags (which can thus correspond to tasks that have not yet been submitted).

// TODO insert illustration f(Ar, Brw, Cr) + ..
// DSM

\subsection StarPUDataManagementLibrary StarPU Data Management Library

Because StarPU schedules tasks at runtime, data transfers have to be
done automatically and ``just-in-time'' between processing units,
relieving application programmers from explicit data transfers.
Moreover, to avoid unnecessary transfers, StarPU keeps data
where it was last needed, even if was modified there, and it
allows multiple copies of the same data to reside at the same time on
several processing units as long as it is not modified.

\section ApplicationTaskification Application Taskification

TODO

// TODO: section describing what taskifying an application means: before
// porting to StarPU, turn the program into:
// "pure" functions, which only access data from their passed parameters
// a main function which just calls these pure functions
// and then it's trivial to use StarPU or any other kind of task-based library:
// simply replace calling the function with submitting a task.

\section Glossary Glossary

A \b codelet records pointers to various implementations of the same
theoretical function.

A <b>memory node</b> can be either the main RAM, GPU-embedded memory or a disk memory.

A \b bus is a link between memory nodes.

A <b>data handle</b> keeps track of replicates of the same data (\b registered by the
application) over various memory nodes. The data management library manages to
keep them coherent.

The \b home memory node of a data handle is the memory node from which the data
was registered (usually the main memory node).

A \b task represents a scheduled execution of a codelet on some data handles.

A \b tag is a rendez-vous point. Tasks typically have their own tag, and can
depend on other tags. The value is chosen by the application.

A \b worker execute tasks. There is typically one per CPU computation core and
one per accelerator (for which a whole CPU core is dedicated).

A \b driver drives a given kind of workers. There are currently CPU, CUDA,
and OpenCL drivers. They usually start several workers to actually drive
them.

A <b>performance model</b> is a (dynamic or static) model of the performance of a
given codelet. Codelets can have execution time performance model as well as
energy consumption performance models.

A data \b interface describes the layout of the data: for a vector, a pointer
for the start, the number of elements and the size of elements ; for a matrix, a
pointer for the start, the number of elements per row, the offset between rows,
and the size of each element ; etc. To access their data, codelet functions are
given interfaces for the local memory node replicates of the data handles of the
scheduled task.

\b Partitioning data means dividing the data of a given data handle (called
\b father) into a series of \b children data handles which designate various
portions of the former.

A \b filter is the function which computes children data handles from a father
data handle, and thus describes how the partitioning should be done (horizontal,
vertical, etc.)

\b Acquiring a data handle can be done from the main application, to safely
access the data of a data handle from its home node, without having to
unregister it.


\section ResearchPapers Research Papers

Research papers about StarPU can be found at
http://starpu.gforge.inria.fr/publications/.

A good overview is available in the research report at
http://hal.archives-ouvertes.fr/inria-00467677.

\section StarPUApplications StarPU Applications

You can first have a look at the chapters \ref BasicExamples and \ref AdvancedExamples.
A tutorial is also installed in the directory <c>share/doc/starpu/tutorial/</c>.

Many examples are also available in the StarPU sources in the directory
<c>examples/</c>. Simple examples include:

<dl>
<dt> <c>incrementer/</c> </dt>
<dd> Trivial incrementation test. </dd>

<dt> <c>basic_examples/</c> </dt>
<dd>
        Simple documented Hello world and vector/scalar product (as
        shown in \ref BasicExamples), matrix
        product examples (as shown in \ref PerformanceModelExample), an example using the blocked matrix data
        interface, an example using the variable data interface, and an example
        using different formats on CPUs and GPUs.
</dd>

<dt> <c>matvecmult/</c></dt>
<dd>
    OpenCL example from NVidia, adapted to StarPU.
</dd>

<dt> <c>axpy/</c></dt>
<dd>
    AXPY CUBLAS operation adapted to StarPU.
</dd>

<dt> <c>native_fortran/</c> </dt>
<dd>
    Example of using StarPU's native Fortran support.
</dd>

<dt> <c>fortran90/</c> </dt>
<dd>
    Example of Fortran 90 bindings, using C marshalling wrappers.
</dd>

<dt> <c>fortran/</c> </dt>
<dd>
    Example of Fortran 77 bindings, using C marshalling wrappers.
</dd>
</dl>

More advanced examples include:

<dl>
<dt><c>filters/</c></dt>
<dd>
    Examples using filters, as shown in \ref PartitioningData.
</dd>

<dt><c>lu/</c></dt>
<dd>
    LU matrix factorization, see for instance <c>xlu_implicit.c</c>
</dd>

<dt><c>cholesky/</c></dt>
<dd>
    Cholesky matrix factorization, see for instance <c>cholesky_implicit.c</c>.
</dd>
</dl>

\section FurtherReading Further Reading

The documentation chapters include

<ul>
<li> Part 1: StarPU Basics
<ul>
<li> \ref BuildingAndInstallingStarPU
<li> \ref BasicExamples
</ul>
<li> Part 2: StarPU Quick Programming Guide
<ul>
<li> \ref AdvancedExamples
<li> \ref CheckListWhenPerformanceAreNotThere
</ul>
<li> Part 3: StarPU Inside
<ul>
<li> \ref TasksInStarPU
<li> \ref DataManagement
<li> \ref Scheduling
<li> \ref SchedulingContexts
<li> \ref SchedulingContextHypervisor
<li> \ref ModularizedScheduler
<li> \ref DebuggingTools
<li> \ref OnlinePerformanceTools
<li> \ref OfflinePerformanceTools
<li> \ref FrequentlyAskedQuestions
</ul>
<li> Part 4: StarPU Extensions
<ul>
<li> \ref OutOfCore
<li> \ref MPISupport
<li> \ref FFTSupport
<li> \ref MICSCCSupport
<li> \ref cExtensions
<li> \ref NativeFortranSupport
<li> \ref SOCLOpenclExtensions
<li> \ref SimGridSupport
<li> \ref OpenMPRuntimeSupport
<li> \ref ClusteringAMachine
</ul>
<li> Part 5: StarPU Reference API
<ul>
<li> \ref ExecutionConfigurationThroughEnvironmentVariables
<li> \ref CompilationConfiguration
<li> \ref ModuleDocumentation
<li> \ref FileDocumentation
<li> \ref deprecated
</ul>
<li> Part: Appendix
<ul>
<li> \ref FullSourceCodeVectorScal
<li> \ref GNUFreeDocumentationLicense
</ul>
</ul>

Make sure to have had a look at those too!

*/