Pixel Bender - Home - AMDdeveloper.amd.com/wordpress/media/2013/06/2903_2_final.pdfDecide if a Pixel Bender node runs on the GPU or CPU Order the execution of each node Drives Optimization

© 2011 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential.

Kevin Goldsmith | Senior Engineering Manager

Bob Archer | Senior Computer Scientist

Pixel Bender


What is Pixel Bender?

A domain specific kernel and graph language for image and video

processing designed to efficiently target current and future

heterogeneous hardware.

2


Pixel Bender Kernels & Graphs

3

<languageVersion : 1.0;>

kernel pixelate_sample

<

namespace : "AIF Test";

vendor : "Adobe";

version : 1;>

{

parameter int dimension;

input image4 inputImage;

output pixel4 outputPixel;

void evaluatePixel()

{

float dimAsFloat = float(dimension);

float2 sc = floor(outCoord() /

float2(dimAsFloat, dimAsFloat));

sc *= dimAsFloat;

outputPixel = sampleNearest(inputImage, sc);

}

}

<?xml version="1.0" encoding="utf-8"?>

<graph name = "PixelateGraph"xmlns="http://ns.adobe.com/PixelBenderGraph/1.0">

<metadata name = "namespace" value = "AIF"/>

<metadata name = "vendor" value = "Adobe Systems" />

<metadata name = "version" type = "int" value = "1" />

<parameter type = "int" name = "dimension" >

<metadata name = "defaultValue" type = "int" value = "1" />

<metadata name = "minValue" type = "int" value = "1" />

<metadata name = "maxValue" type = "int" value = "100" />

</parameter>

<inputImage type = "image4" name = "inputImage" />

<outputImage type = "image4" name = "outputImage" />

<kernel>

<![CDATA[


kernel Pixelate

<

namespace:"AIF";

vendor:"Adobe Systems";

version:1;

>

{

…

}

]]>

</kernel>

<node id = "pixelateFilter" name ="Pixelate" namespace = "AIF" vendor = "Adobe Systems"

version ="1" clientID ="ADBE Pixelate" >

<evaluateParameters>

<![CDATA[

void evaluateParameters()

{

pixelateFilter::dimension = dimension;

}

]]>

</evaluateParameters>

</node>



<connect fromImage = "inputImage" toNode = "pixelateFilter" toInput = "src" />

<connect fromNode = "pixelateFilter" fromOutput = "dst" toImage = "outputImage" />

</graph>



4



<


vendor : "Adobe";

version : 1;>

{





{




sc *= dimAsFloat;


}

}










</parameter>



<kernel>

<![CDATA[


kernel Pixelate

<

namespace:"AIF";


version:1;

>

{

…

}

]]>

</kernel>




<![CDATA[


{


}

]]>


</node>




</graph>



5



<


vendor : "Adobe";

version : 1;>

{





{




sc *= dimAsFloat;


}

}










</parameter>



<kernel>

<![CDATA[


kernel Pixelate

<

namespace:"AIF";


version:1;

>

{

…

}

]]>

</kernel>




<![CDATA[


{


}

]]>


</node>




</graph>



6



<


vendor : "Adobe";

version : 1;>

{





{




sc *= dimAsFloat;


}

}










</parameter>



<kernel>

<![CDATA[


kernel Pixelate

<

namespace:"AIF";


version:1;

>

{

…

}

]]>

</kernel>




<![CDATA[


{


}

]]>


</node>




</graph>

High Pass Texturize

Saturate

Blend Vacation



7



<


vendor : "Adobe";

version : 1;>

{





{




sc *= dimAsFloat;


}

}










</parameter>



<kernel>

<![CDATA[


kernel Pixelate

<

namespace:"AIF";


version:1;

>

{

…

}

]]>

</kernel>




<![CDATA[


{


}

]]>


</node>




</graph>

•Explicitly Data Parallel

•Explicitly Lock Free

•Implicitly Vectorizable

•Contains Optimization Hints

•Region Reasoning

•Parameter Ranges

•Supports per-frame

functions

•JIT

•x86 with SSE

•GLSL

•Implicitly Task Parallel

•Contains Optimization Hints

•Region Reasoning

•Parameter Ranges

•Parameter Use

•Supports per frame and internal

graph logic

•Language/Hardware Agnostic

•Enforces the programming Model


Pixel Bender Runtime

Run-time scheduling

Decide if a Pixel Bender node runs on the GPU or CPU

Order the execution of each node

Drives Optimization

Runs region reasoning pass

Concatenate adjacent Pixel Bender nodes

Fix parameter values for constant folding and range propagation

Caches intra-graph frames when possible to reduce computation

Allows host application to aid optimization

Lock parameter values

Lock inputs

Interfaces with host application resource management

Application can force CPU or GPU execution

Application can provide memory or threading management

8



Run-time scheduling



Drives Optimization







Lock inputs




9



Run-time scheduling



Drives Optimization







Lock inputs




10



Run-time scheduling



Drives Optimization







Lock inputs




11


Demo

12


Lessons Learned (so far)

A DSL allows you to limit the language for optimization purposes while

preserving programmability

Just-in-time Compilation allows you not only to use runtime information

for optimization purposes, but also allows you to future-proof algorithms

against later hardware architectures

Developers will learn new programming languages if they give them an

ability to do something new, or make something significantly easier

A language with only a text representation creates some nice network

and community effects (for a time)

13












14












15












16


Warning.

Speculation ahead.

17


How do we make the most of the

hardware we have available?

18


Writing multi threaded code is like

juggling chainsaws; amazing when it

works and truly sucky when it doesn’t.

Andrew Wulf

19


Gflop / $

20


Gflop / $

Gflop / watt

21


Gflop / $

Gflop / watt

Gflop / developer-hour

22


What have we got?

23


What have we got?

DSL for image processing (Pixel Bender)

24


What have we got?


JIT compiler / optimizer

25


What have we got?



Runtime – enforces programming model, allows graphs

26


What have we got?




High performance on parallel hardware (using OpenGL)

27


What have we got?





Future proof against new hardware

28


What have we got?






Cross platform

29


What cross platform means to Adobe

Mac

Windows

AMD

Intel

nVidia

CPU

GPU

Mobile hardware

30


What have we got?






Cross platform

31


What have we got?






Cross platform

OpenCL

32


Why not just switch entirely to OpenCL?

33



Pixel Bender has certain advantages:

Ease of writing

34




Ease of writing

Optimization opportunities

35




Ease of writing


Race free by construction

36




Ease of writing


Race free by construction

Parallelism for the masses

37


Warning.

Increased speculation

ahead.

38


Pixel Bender has certain disadvantages:

One fixed programming model

Focuses on image processing

Not general enough

39


Framework

40

Inputs Framework Outputs


Framework

41

Inputs Framework Outputs


Michael McCool’s parallel programming patterns

Gather

Map

Reduce

Superscalar sequences

Pipeline

Nesting

Scans

Recurrences

Search

Subdivision

Stencil

Scatter

Pack

Selection

Partition

Expand

42


What are we going to do about it?

43


Internal customers

44


External customers

45


References

Pixel Bender

http://www.adobe.com/devnet/pixelbender.html

http://blogs.adobe.com/pixel-bender/

http://blogs.adobe.com/kevin-goldsmith/tag/pixel-bender-2

Structured Parallel Programming with Deterministic Patterns

Michael D. McCool http://software.intel.com/file/27160

Patterns for Parallel Programming

Mattson, Sanders & Massingill

46

http://www.adobe.com/devnet/pixelbender.html











http://software.intel.com/file/27160


© 2011 Adobe Systems Incorporated. All Rights Reserved. Adobe Confidential. 4

8

Pixel Bender programming model

Write a function that produces a

single output pixel


9



0



1



2




Why Pixel Bender?

2001 programmable shading units on the GPU

2005 dual core CPUs

2005 AIF team formed

2006 quad core CPUs

2006 AMD announces CTM

2008 OpenCL

2011 fusion chips

54


Pixel Bender History

2001 programmable shading units on the GPU

2005 dual core CPUs

2005 AIF team formed

2006 quad core CPUs

2006 AMD announces CTM

2007 After Effects CS3 replaces GLSL code with Pixel Bender

2007 Flash Player 10 Announced including Pixel Bender support

2008 OpenCL

2008 After Effects CS4 - first CPU implementation, 3rd party plug-ins, Pixel Bender Graph language announced

2008 Photoshop CS4 – Pixel Bender plug-in released

2010 Pixel Bender commercial plug-ins start to appear en masse

2010 Flash Player Molehill Announced – Pixel Bender 3D

2011 fusion chips

55


Disclaimer & Attribution

The information presented in this document is for informational purposes only and may contain technical inaccuracies, omissions and typographical errors.

The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to product and roadmap changes, component and motherboard version changes, new model and/or product releases, product differences between differing manufacturers, software changes, BIOS flashes, firmware upgrades, or the like. There is no obligation to update or otherwise correct or revise this information. However, we reserve the right to revise this information and to make changes from time to time to the content hereof without obligation to notify any person of such revisions or changes.

NO REPRESENTATIONS OR WARRANTIES ARE MADE WITH RESPECT TO THE CONTENTS HEREOF AND NO RESPONSIBILITY IS ASSUMED FOR ANY INACCURACIES, ERRORS OR OMISSIONS THAT MAY APPEAR IN THIS INFORMATION.

ALL IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE ARE EXPRESSLY DISCLAIMED. IN NO EVENT WILL ANY LIABILITY TO ANY PERSON BE INCURRED FOR ANY DIRECT, INDIRECT, SPECIAL OR OTHER CONSEQUENTIAL DAMAGES ARISING FROM THE USE OF ANY INFORMATION CONTAINED HEREIN, EVEN IF EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

AMD, the AMD arrow logo, and combinations thereof are trademarks of Advanced Micro Devices, Inc. All other names used in this presentation are for informational purposes only and may be trademarks of their respective owners.

The contents of this presentation were provided by individual(s) and/or company listed on the title page. The information and opinions presented in this presentation may not represent AMD’s positions, strategies or opinions. Unless explicitly stated, AMD is not responsible for the content herein and no endorsements are implied

56

Documents

Pixel Bender - Home - AMDdeveloper.amd.com/wordpress/media/2013/06/2903_2_final.pdfDecide if a Pixel Bender node runs on the GPU or CPU Order the execution of each node Drives Optimization