VizWorld.com

Whether you’re at GTC this week or not, you’ve no doubt heard the amazing claims of performance boosts possible with using not just CUDA, but any GPU-acceleration system.  However, you can’t just recompile your code and expect massive boosts, you have to dig deep and invest the time and tools to completely rebuild your algorithms for maximum parallelism and concurrency.  This week’s recommended resource is a guide to help you do just that: Clay Breshears’ “The Art of Concurrency“.

If you’re looking to take full advantage of multi-core processors with concurrent programming, this practical book provides the knowledge and hands-on experience you need. The Art of Concurrency is one of the few resources to focus on implementing algorithms in the shared-memory model of multi-core processors, rather than just theoretical models or distributed-memory architectures. The book provides detailed explanations and usable samples to help you transform algorithms from serial to parallel code, along with advice and analysis for avoiding mistakes that programmers typically make when first attempting these computations.

Written by an Intel engineer with over two decades of parallel and concurrent programming experience, this book will help you:

• Understand parallelism and concurrency
• Explore differences between programming for shared-memory and distributed-memory
• Learn guidelines for designing multithreaded applications, including testing and tuning
• Discover how to make best use of different threading libraries, including Windows threads, POSIX threads, OpenMP, and Intel Threading Building Blocks
• Explore how to implement concurrent algorithms that involve sorting, searching, graphs, and other practical computations

The Art of Concurrency shows you how to keep algorithms scalable to take advantage of new processors with even more cores. For developing parallel code algorithms for concurrent programming, this book is a must.

Of course, if you already know everything there is to know about Parallel Algorithms but just need a refresher on CUDA, you can’t go wrong with the last Recommended Resource: CUDA By Example.

This book and many others is available in the Vizworld Store.

Science feature, resource of the week

If you were reading our LiveBlog of the GTC2010 Keynote Speech, then you saw the note about the Product Roadmap. NVidia has historically been pretty secretive with roadmaps, particularly with the press, but this time Jen-Hsun threw up a single slide announcing not only the names of their upcoming 2 new products, but estimated performance figures and dates.

The two new products:

• Kepler – To be released sometime in 2011, 28nm process.  Estimated double-precision Gigaflops performance of 4-6GFlops per watt.
• Maxwell – To be released sometime in 2013, 22nm process.  Estimated double-precision Gigaflops performance of 15-16 GFlops per watt, making is about 16x better than the Fermi-driven cards.

How can they expect to hit these unbelievable numbers?  First off, they’re going to work on reducing the process size (Fermi is at 40nm) so simply cutting down to 28nm would allow a significantly quantity of extra transistors to be added.  However, that’s not the only thing that will have to happen.  Beyond that, NVidia wouldn’t say much more than “architectural changes” similar to what happened with Fermi.

I was, personally, very glad to see NVidia presenting figures in Performance Per Watt, as in Jen-Hsun’s words:

Math is Free.  Transistors are Free.  Power is expensive.  Performance Per Watt = Performance.

And this translates through their entire product portfolio.  The changes that will enable this kind of performance on the high-end will also show up in the consumer-side (GeForce) as improved gaming performance, better PhysX, and more real-time raytracing, and it will show up on the mobile side as improved Tegra chipsets (CUDA in the power of your hand??!).

Hardware feature, gtc, nvidia

On stage at GTC2010, Ken Pimental (Autodesk), Micheal Kaplan (mental images) and Jen-Hsun Huang (NVidia) stood together to announce the next generation rendering technology they’ve integrated.  They showed a simple scene of a few chairs around a table that had been rendering the scene for over an hour, and was approximately 10% in when they were on stage.   Mainly, the delay would due to the fast that the scene has no direct lighting, all lighting was indirect through the set of double-glass windows in the back and then reflected off the various elements in the scene.  They then paused it to show the various Render panels, showing the details of Final Gather and illumination settings, about how they have had to include many options to control the various shortcuts.

The iRay Control Panel.

Starting next week, subscription owners will be able to download iRay, mental image’s real-time GPU-accelerated ray-tracing solution, for use directly within 3dsMax as a new renderer.    The result is a far simpler control panel, far faster renders, and vastly more accurate images at the end.  On the right, you can see the control panel (click for larger-size), and you can see how trivial the new setup is.

After that, they clicked the “Render” button.  It took approximately 10s to translate and pad the scene into the proper formats, then within seconds a near-complete render was visible.  iRay is an interative renderer so the longer you let it go, the better the result will be, however the initial result is usually good-enough (to tell that something is wrong, needs to be corrected, etc) making the iterative process MUCH faster.

The same scene in iRay, 5seconds in.

This will be available to 3dsMax Subscription members next week.

Next, they showed the “cloud rendering” capabilities on stage, run by Peer 1 hosting, showing how the iray rendering (on 32 GPUs) offers interactive real-time rendering.  In addition to simply rendering the image faster, it enables the user to actually “walk” around the scene and see it rendered interactively as they use it. Via a custom web interface, they could then place furniture and art within the room, all rendered realtime.  In addition, since there was no direct lighting they could change the time of day to see completely different lighting environments.

iRay on the Web

The web-interface is all research, and I’ve seen it several times before, but the “official” presentation is new.

Graphics, Science autodesk, feature, gtc, mental images, nvidia

I’m settled into the press pit at the NVidia GPU Technology Conference Keynote hall, where the screens are unreasonably large and the microphones are deafeningly loud.  Currently listening to “Uprising” by Muse, and waiting for it to begin.  I’ll be updating this post as new information comes along, so stay tuned!

• 8:42 – I don’t know who makes NVidia’s playlists but I need to find them.. Now listening to “Electric Eel” by MGMT.  Seems several press photographers are using HDR cameras, listening to the rapid-fire shutterclicks.
• 8:44 – Don’t forget you can watch along (if you have the ability) at NVidia’s livestream of the keynote.
• 8:45 – Seems everyone gets a pair of Dolby3D Passive Polarized glasses upon entry.  I guess 3D will be on the agenda.
• 8:50 – Beginning to notice a theme.. “Uprising”, “Keep on rocking in the free world”… Hrrm.
• 8:56 – Before we begin, enjoy some “really kewl demos” in 3D..
• Technical Snafu there.. sorry folks. Had to disable the Twitter Tracker temporarily.
• 9:11 – Ubisoft HAWX2 Demo, in 3D.  NVidia has 3 focus areas: Professional (Quadry), Computational (Tesla), Personal (Tegra/GeForce).

Remainder after the break.

Read more…

Hardware feature, gtc, nvidia

Microsoft announced today that it is releasing Windows HPC Server 2008 R2. This release of Windows HPC Server 2008 R2 is part of Microsoft’s Technical Computing Initiative. Windows HPC Server has the ability to scale out to thousands of nodes.

What is new in this release?

First, Microsoft is announcing the ability of Windows HPC Server to distribute Excel worksheets across a cluster. For companies with large worksheets, this could potentially reduce the amount of time that it takes to compute an Excel worksheet. It does not need to be said, but this is not your traditional HPC problem.

Second, Microsoft is announcing the ability of Windows HPC Server to integrate Windows 7 workstations into the compute cluster. I can easily imagine that some companies may have idle Windows 7 workstations at night. That idle time could be used to solve problems overnight. In fact, at my first job with McDonnell Douglas (now part of Boeing), one of my colleagues used PVM to run on SGI workstations overnight. This is something that Windows HPC Server can now do with Windows 7.

Third, Microsoft is announcing the ability of Windows HPC Server to integrate with Windows Azure. Windows Azure is Microsoft’s cloud platform offering. Now I am pretty skeptical of clouds for simple reason: “What about the security of my data?” If I am a corporation, I do not want my competitors to gain access to my data in the cloud. Microsoft assures me that they have 9 layers of security to prevent this. For it to really take off for HPC computing, the public cloud will have to meet the local security policies of the corporation or government.Update: Microsoft “will release an update to Windows HPC Server that allows customers to provision and manage HPC nodes in Windows Azure from within on-premises server clusters in the near future.”

You can read the press release after the break.

Read more…

Hardware feature, microsoft

Here’s your collection of the posters presented at the NVidia GPU Technology conference 2010.  I tried my best to take the pictures at resolution high enough that you can still read them.  I have them at resolution even higher still, but filesize limits began to become a problem.  If there’s any particular poster you want in higher resolution, post in the comments and I’ll share them individually!

Read more…

Science feature, gtc, posters

This week’s recommended resource is for anyone gearing up for NVidia’s GPU Technology Conference at the end of this month, and comes straight from two senior developers in the CUDA software platform team, the recently released CUDA By Example.

CUDA by Example, written by two senior members of the CUDA software platform team, shows programmers how to employ this new technology. The authors introduce each area of CUDA development through working examples. After a concise introduction to the CUDA platform and architecture, as well as a quick-start guide to CUDA C, the book details the techniques and trade-offs associated with each key CUDA feature. You’ll discover when to use each CUDA C extension and how to write CUDA software that delivers truly outstanding performance.

Major topics covered include

• Parallel programming
• Thread cooperation
• Constant memory and events
• Texture memory
• Graphics interoperability
• Atomics
• Streams
• CUDA C on multiple GPUs
• Advanced atomics
• Additional CUDA resources

All the CUDA software tools you’ll need are freely available for download from NVIDIA.
http://developer.nvidia.com/object/cuda-by-example.html

This book is actually the foundation (I’m told) of a recent Webinar series on CUDA, sponsored by Nvidia.  This book, and many others, is available in the VizWorld store.

Science cuda, feature, nvidia, resource of the week

This week’s resource aims to teach you everything you ever wanted to know (probably much more, in fact) about High Dynamic Range Imaging and its use in Image Based Lighting.  This is actually the second edition of the original High Dynamic Range Imaging text,High Dynamic Range Imaging, Second Edition: Acquisition, Display, and Image-Based Lighting.

High Dynamic Range Imaging was the first book to describe this exciting new field that is transforming the media and entertainment industries. The second edition brings a significant update, adding chapters on high dynamic range image capture (hardware and software), display devices, as well as image difference metrics and video. All existing chapters have been updated to reflect the current state of the art, ensuring the book’s leading position as a reference text for those working with images, whether it is for computer graphics, film, video, photography, or lighting design.

• Up-to-date revision of the “BIBLE” of High Dynamic Range Imaging
• New material includes chapters on High Dynamic Range Video Encoding, High Dynamic Range Image Encoding, and High Dynamic Range Display Devices
• Invaluable reference for anyone serious about computer graphics, interactive entertainment, and photography/imaging
• Written by the inventors and initial implementors of High Dynamic Range Imaging
• Covers the basic concepts (including just enough about human vision to explain why HDR images are necessary), image capture, image encoding (not as easy as it sounds), file formats, display techniques, tone mapping for lower dynamic range display (FAR from easy), and the use of HDR images and calculations in 3D rendering (which is very cool, even if you aren’t working in 3D)
• Range and depth of coverage is good for the knowledgeable researcher as well as those who are just starting to learn about High Dynamic Range imaging

Graphics feature, hdri, resource of the week

This week’s recommended resource is the latest (5th) edition of the OpenGL Superbible from Richard Wright and Addison-Wesley publishing.

OpenGL® SuperBible, Fifth Edition is the definitive programmer’s guide, tutorial, and reference for the world’s leading 3D API for real-time computer graphics, OpenGL 3.3. The best all-around introduction to OpenGL for developers at all levels of experience, it clearly explains both the API and essential associated programming concepts. Readers will find up-to-date, hands-on guidance on all facets of modern OpenGL development, including transformations, texture mapping, shaders, advanced buffers, geometry management, and much more. Fully revised to reflect ARB’s latest official specification (3.3), this edition also contains a new start-to-finish tutorial on OpenGL for the iPhone, iPod touch, and iPad.

Coverage includes

• A practical introduction to the essentials of real-time 3D graphics
• Core OpenGL 3.3 techniques for rendering, transformations, and texturing
• Writing your own shaders, with examples to get you started
• Cross-platform OpenGL: Windows (including Windows 7), Mac OS X, GNU/Linux, UNIX, and embedded systems
• OpenGL programming for iPhone, iPod touch, and iPad: step-by-step guidance and complete example programs
• Advanced buffer techniques, including full-definition rendering with floating point buffers and textures
• Fragment operations: controlling the end of the graphics pipeline
• Advanced shader usage and geometry management
• A fully updated API reference, now based on the official ARB (Core) OpenGL 3.3 manual pages
• New bonus materials and sample code on a companion Web site, www.starstonesoftware.com/OpenGL

I’ve been looking over this book for the last few days (they were kind enough to provide me a review copy) and I’m impressed.  While it, admittedly, doesn’t mention anything about the new OpenGL4.0 or 4.1 specs, it does a great job covering OpenGL3.3.  It opens with all of the basics of geometry, matrix math, and texturing, before moving into buffer objects, vertex and pixel shaders, and geometry queries.  It even gets platform specific at the end, detailing things you need to know about the specifics of using OpenGL 3.3 on Windows, OSX, or Linux, and contains a chapter on OpenGL ES for mobile devices, including a chapter or two specifically for the iPhone.

I highly recommend it, and you can find this book and many others in the VizWorld Store.

Graphics, Science feature, opengl, resource of the week

At SIGGRAPH2010, Autodesk was advertising a collection of new ROI studies done by Pfeiffer Consulting, showing the vast impact newer versions of 3dsMax and Maya have had on their users.  One frequent argument from managers and people not intimately familiar with the products is what changed between 3dsMax 2009 and 2008 that justifies the expense?  It’s still a modeling, animation, and rendering package, so what justifies the outlay of cash?

The new reports attempt to answer this by analyzing the impressive improved performance offered in the newer versions of the Autodesk products, and benchmarks several frequently used features.  They present the results like so, showing the improvement between 3dsMax 2008 and 3dsMax 2011.

After analyzing several aspects of the software, focusing heavily on Autodesk’s admitted “big improvements” like the new shader manager, the Quadrify functions, and improved rigging tools, they then take the several areas they benchmarked and calculate how many times a day this happens in an animators life, and how much time is saved over the course of a day, week, and year.  Then, for the most substantial impact to the bean counters who are typically reponsible for signing off on the purchases, they look at the amount of money you “save” over the course of a year by upgrading.  The results:

• At $100/hr, an animator could “lose” almost $16,000/year in 3dsMax 2008
• At $100/hr, an animator could “lose” almost $20,000/year in Maya 8.5

The studies are well done, and contain a wealth of information about the improvements Autodesk has made in the last several years, both through routine optimization of the code and the new “Project Excalibur” initiative they’ve been working on.

“Our customers are constantly balancing creativity and innovation with the need to meet tight production deadlines and budgets,” said Stig Gruman, Autodesk vice president, digital entertainment. “As a result, Autodesk has focused on developing both innovative creative tools as well as on under-the-hood architectural changes to improve the performance of Maya and 3ds Max. The cumulative impact of the improvements assessed in the Pfeiffer studies on modern production pipelines is significant. The Pfeiffer studies help illustrate how productivity gains in everyday operations can lead to savings of thousands of dollars per year.”

Personally, I think information like this is invaluable for not just Autodesk products, but all graphic design products.  As these products continue their annual release cycle they risk running into situations similar to Microsoft’s Office suite where users (and professionals) simply stop upgrading since there is no perceivable reason to.  Photoshop CS5 still touches up photos, just like CS2, so why upgrade?  For most consumers, it doesn’t matter, but for professionals these numbers are critical, justifying not only why they should upgrade but quantifying the benefits of such.

However, I do take some issue with why they chose to benchmark their new products (2011) against such old versions.  Why not compare 3dsMax2011 against 2010, instead of the 3-year old 2008?  Why not compare Maya 2011 against 2010, instead of the 2007 version.  Of course this makes for the most impressive results in the ROI study, but is anyone out there in the professional space still using Maya 8.5?

You can read a full press release about the ROI studies here, or read the individual reports:

• Autodesk 3ds Max 2011 Market Perspectives, Productivity and Return on Investment
• Autodesk Maya 2011 Market Perspectives, Productivity and Return on Investment
• Autodesk 3ds Max 2011: Productivity Benchmarks
• Autodesk Maya 2011: Productivity Benchmarks

If you have a question or concern about the study, post in the comments!

Graphics autodesk, feature, siggraph