The obvious second part of 
any year-end wrap-up, right after reflecting on the year gone by, is to look to the future and see what lies ahead. In this post I attempt to decipher the runes, gaze into the crystal ball, and talk to industry contacts to see what we can expect in the upcoming year. I’ll try to skip the obvious (Autodesk will release a new 3dsMax and Maya, Adobe will release Photoshop CS5, etc) and cover some of the bigger and less obvious events that we can expect in the upcoming year.
Read on after the break and post in the comments why you think we’re wrong or right.
NVidia Slips, but Remains #1
The first prediction of 2010 is that NVidia will lose a significant chunk of marketshare to AMD/ATI. Windows7 has been around almost a year (in beta, pre-release, and formal release) and NVidia has yet to release a DirectX11 card. While the number of games and applications that are DirectX11 only is pretty miniscule, hardware providers will be relunctant to ship systems with non-compliant hardware and will instead turn to ATI/AMD cards. In addition, ATI is getting a lot of publicity right now with their continual onslaught of new video cards, and that will continue into 2010. AMD/ATI is enjoying their time in the lead, and it shows.
However, while NVidia will lose ground on the consumer front, it’s enjoying the solo spot of GPU Computing and will continue to for all of 2010. They own CUDA, and OpenCL currently only really works on NVidia hardware. ATI’s implementation is a bit kludgy and buggy, but I’m sure it will improve. Nonetheless, NVidia has the Tesla & Quadroplex systems, favorites among GPGPU developers and the Fermi will only advance their lead.
NVidia ships Fermi “on time”
That’s right, I believe Fermi will make their Q1 2010 release date. On paper, at least. Commitments have already been made for large-scale deployments, with deadlines that they can’t afford to miss. As such, the Fermi cards will begin to ship on schedule in Q1 2010. However, the quantity will be very limited and dedicated to these large deployments, meaning that you won’t actually be able to go out and buy one until late Q2 or, more likely, Q3.
Stereoscopic: Wins in Theaters, Flops at Home
2010 will see more stereoscopic 3d movies, and they will continue to rake in big bucks compared to their 2D counterparts. The difficulty in pirating them, combined with the extra money theaters (and therefore studios) make in selling tickets makes it a no-brainer for the industry.
At home, however, the new Stereoscopic Blu-Ray standard will simply flop. Why, you ask? Several reasons:
- Existing BluRay players probably will not support it, so people interested will have to go buy new BluRay players.
- It requires a 120Hz Television, restricting the market even more.
- They chose active Shutter glasses, which adds a significant cost to get started ($75-$100 for only a pair).
While I’m glad to see stereoscopic making a push to the home, and I understand that they chose the Active Shutter technology because of it’s compatibility with existing systems, I simply don’t think it will catch on in-the home. The exception will be the true hardcore movie-goers who have the several thousand dollar home theater rigs in their house. In that respect, I expect Stereoscopic BluRay to be alot like Laserdisc, toys for the wealthy bachelor.
Update 12/30/2009: Comments on FaceBook have brought up one or two things that probably need further discussion. From the details in this ArsTechnica story, apparently active shutter glasses aren’t really a Requirement. The new codec they are using is backwards compatible with existing bluray players, in the sense that existing players will still be able to watch the movie in 2D. For a 3D effect, you’ll need to get a new player (or possibly apply a Firmware update, as will be the case with the PS3). In the new players, they will decode the left and right images into two separate 1080p streams, and leave it to the TV to determine what to do. Based on current technology, the only widely available option is 120Hz TV’s, which will require active stereoscopic images.
The alternative to this is the new generation of TV’s with built-in support for 3D via passive glasses. Given how long it’s taken for LCD HD TV’s to penetrate the market, these new TV’s will be slow growing and may even flop out of the gate from poor sales. Add to that the niche that BluRay remains (Hit any big-box store and see how small the BluRay section is compared to DVD), I still don’t think we’ll see widespread 3D in the home anytime soon.
Essentially, it breaks down to this: Take the US population..
- What percentage of them care about 3D?
- What percentage of BluRay titles in 2010 will support 3D?
- What percentage of that is willing to buy a new BluRay player?
- What percentage of that is willing to buy A) A new TV, or B) Stereoscopic glasses?
The end result is a pretty small percentage, and not enough to sustain any sizable market I believe.
Real-Time Ray Tracing gets big
For several reasons, 2010 should be the year that real-time raytracing (RTRT) becomes big. In the animation space, GPU accelerated ray-tracing offers faster previews of scenes, reducing development and tinkering time and getting to the final render faster. In the visualization space, ray-tracing is much easier to parallelize across multiple systems in a near-linear fashion, unlike current rasterization/composition methods (like Mesa).
However, real-time raytracing will not become a full replacement for existing rendering solutions. Tools like MentalRay and VRay will not find themselves usurped by real-time raytracing solutions, only enhanced. While RTRT is great for previs and draft runs, final products will still require the extra touches only available from MentalRay, RenderMan, or VRay.
Next Generation Console News
(This one is already beginning) In 2010, we’ll start to hear news of hardware components that will appear in the next generation of gaming consoles. While the consoles themselves won’t actually materialize until 2011 (or 2012), hardware selections for graphics chipsets, resolutions, and framerates will be made this year. I think Sony learned their lesson with the exotic Cell processor, yielding an incredibly powerful console that no one can understand how to code for, and will instead adopt an approach much closer to the Xbox, using standard components.
So that’s my Predictions for 2010.. I know it’s a bit light on Cinema and VFX predictions, but I’m willing to accept input on anyone who cares to make their own and share them!
@andrew Of course, when you’re talking about Computer Hardware, you’ll regret everything “2 years later”. Such is the nature of the industry right now.
@tchavez I never meant to imply otherwise. However, AMD’s current implementation has issues. The poor performance on 4XXX series hardware, lack of any suitable way to intercept the IL/ISA at compile-time, and several other minor annoyances that basically boil down to “It’s the First Release”.
Later releases will surely improve, but for now NVidia has the advantage of having several version of CUDA & OpenCL out in the wild for user feedback.
Actually, OpenCL is supported on both Nivida and AMD hardware. For more information please visit:
http://developer.amd.com/gpu/ATIStreamSDK/Pages/default.aspx
Quote: (I fail to see how PCI Express 3.0 and an Intel 8-core processor will “Destory” NVidia.)
Hmmm… I didn’t say PCI exp 3.0 and Intel 6(8) core CPU will destroy nVidia – all I said was – that $2000-$4000 still raw Teslas that nVidia is so anxious to release right now would be much more suitable for later. And what’s important for now… (dx11 market btw)… this is where they lose big. nVidia’s got potential, but if they continue to screw around, showing middle class users -“Oh, you guys can wait – we have more important things to take care of…” ATI will crash them on it(oh, sorry, they already did), and then, ATI will go after FERMI-like-Tesla technology and design it for faster PCIe slots and something more advanced than lousy 1066Mhz-1333Mhz integrated mem controller – obviously FOR LESS! the companies who will pay 2k-4k for these teslas now, will regret 2 years later(maybe less). so simple. LEARN, nVIDIA, LEARN, BABY, LEARN, WHILE I AM STILL ALIVE!!! 🙂
@andrew I fail to see how PCI Express 3.0 and an Intel 8-core processor will “Destory” NVidia. If anything, PCI Express 3.0 will make the Tesla & Fermi _more_ attractive, reducing the one thing they really suck at : moving data back and forth to main memory.
And an 8-core Intel CPU still pales to a 512-core Fermi on easily paralleled tasks.
Bullshit. real 100% bullshit. And note: the writer of this article with his coool prediction abilities and sure super-duper mind flexibilities forgets to mention that when PCI express 3.0 boards will hit the stores in 2011, and Intel will relese a 10Gt/s 6(8) core cpu, with improved mem. controller hopefully. and year later… the entire series of $2000-$4000 nVidia’s tesla won’t be worth a broken nickle. Why? Because ATI won’t be just sitting down not doing much… It was a big mistake to give full time to those teslas, knowing that DX11 is coming out and that new upcoming six-core Intel’s CPU is for 1366 (means old, retarded mem. controller), PCI express 3.0 is not out yet… blah blah blah… Very poor strategy, nVidia.