Back in 2008 I took a job with Intel. Intel was starting to ramp up its Larrabee effort and was looking to hire some knowledgeable graphics engineers who could also talk to game developers about *other* things like threading, performance, and optimization. This kept me pretty busy with not enough time to talk about graphics, optimization. And there’s no though about talking about Larrabee since that was not my group. Unfortunately Larrabee wasn’t ready in time so that aspect of my job, while quite a lot of fun, won’t be ready for public consumption for a while. On the other hand, the other aspects of my job were chugging along and I get to do conference talks of these aspects as well, so expect to see some information on multithreading using tasks, using performance tools, programming on Intel’s integrated graphics cores (which really don’t suck anymore), and something new for me, data parallelism though SSE/AVX programming. Since this stuff isn’t as secret as Larrabee, I’m reactivating and reworking my old website into a blog format, since that what is was way back before they had anything called a blog.
It’s no great secret that Intel has been eyeing the discreet graphics market. Intel typically owns about 30-40% of the desktop graphics market, but that’s strictly integrated (and hence – usually considered underpowered) graphics. ATI and Nvidia own most of the other part of the market and anyone who’s interested in playing 3D games wouldn’t consider using an integrated graphics solution if they could help it. Apparently the acquisition of ATI by rival AMD has set Intel to aggressively start talking about discreet graphics. In fact Intel has recently started aggressively hiring engineers (both software and hardware) for their Visual Computing Group. From the recruitment copy:
Join us as we focus on strengthening our leadership in integrated and high-throughput graphics and gaming experiences by developing innovative processing products based on a many-core architecture. We’re looking for engineers, developers, and architects who share our vision and understand what can happen when serious skills and vast resources join forces.
It would seem that Intel is pushing a multi-core architecture for the 2008-2009 timeframe. Given Intel’s manufacturing chops they could, if they set their mind to it, make a pretty deep impression on the discreet graphics market. Given that timeframe you’re looking at a 10x to 20x performance boost over the current top GPU, an Nvidia G80. Even if Intel makes a few missteps and produces something that’s underpowered compared to the best from ATI or Nvidia, they would still probably be competitive on price alone.
Or perhaps they could be doing something to surprise us all? Intel recently did something pretty smart, they hired Daniel Pohl, recent Erlangern University graduate. Daniel coauthored a paper – Realtime ray tracing for current and future games in which the case is made that the traditional hardware rendering pipeline – i.e. object built up independently of each other with depth created through the use of a z-buffer, and all object-object visual interactions (shadows, reflections, etc.) having to be added on – has just about reached the end of its lifetime. Daniel is pushing raytracing instead. Raytracing lets you build up complex scenes in which all the lighting, shadowing, transparency, reflection, caustics, etc. are all handled by the ray tracer. Granted, that the framerates that Daniel gets on the custom raytracer board were about 10% of current consumer level boards, but raytracing will be the next big step in graphics architecture, not a kludge like the doomed Talisman architecture. Raytracing really is the way to render scenes. Maybe Intel will be the first?
At the GDC, graphics chip maker NVIDIA announced they are releasing a bunch of updated and new tools. The tool upgrades are: FX Composer 2, PerfHUD 5, ShaderPerf 2. They are also releasing a new GPU-accelerated texture tool, plus a Shader Library. The most interesting toolkit is the DX10 SDK. Targeted towards the GeForce 8 series of GPU (the only GPU that can run D3D10 so far), it’s a collection of samples for both OpenGL and DirectX, executables and source, that demonstrate and showcase DX10 features. The installer looks and acts like the Microsoft DX Sampler.
The DirectX collection makes up the bulk of the samples, while the OpenGL side is a bit thin. The samples include Rain, Smoke, Fur, Shadows, cloth simulation, Render Target usage, etc. The Direct3D SDK is 256MB while the OpenGL SDK is 45MB.
To compile the source you’ll need Microsoft Visual Studio 2005 plus have the Feb. 2007 DirectX SDK installed (which you can get here.) if you want to compile the DX samples.
If you actually want to run the code you’ll need a DirectX10 video card – which currently means an NVIDIA GeForce 8. There are videos of the programs so even if you don’t have a DX10 video card you can still see the programs running.
Data released on Monday by Mercury Research shows that ATI edged out NVIDIA in sales for Q4 2003. The numbers were Intel 31.7%, ATI 24.9%, and NVIDIA 24.7%. ATI increased at Intel & NVIDIA expense. Most of the change was in add-in cards (vs. integrated chip-sets), where ATI has handily been beating NVIDIA. This was quickly followed by Goldman Sachs analyst Andrew Root questioning if NVIDIA can meet is revenue estimates, though he didn’t change his estimates for the company. NVIDIA’s stock price has dropped about 10% this year while ATI’s has increased about the same amount. NVIDIA will host a conference call to discuss its financial results for the fourth quarter and the fiscal year ending January 25, 2004 on February 12, 2004 at 2:00 PM, Pacific Time. The Company’s prepared remarks will be followed by a question and answer period, which will be limited to questions from analysts and institutional investors. To listen to the conference call, dial 706-679-0543; no password is required. The conference call will also be webcast at www.nvidia.com and www.streetevents.com.
Microsoft officials have posted slightly better than expected figures for the company’s first quarter results. Unusually sales were boosted from outside of the company’s core business, in particular servers and MSN. For the period ended September 30 sales rose 6% from the same period last year to $8.22 billion. Excluding employee stock-based compensation, which have only been reported this quarter, the company earned 30 cents a share – one cent higher than analysts had expected. Net income also increased to $2.61 billion, up from $2.04 billion last year.
Importantly the company saw a 20% revenue growth in its home and entertainment unit, which includes the Xbox, with $581 million in revenue generated. Nevertheless shares in the company still fell once the figures were announced, due to news that fewer large corporate contracts for software had been agreed than expected. This fall in contracts is largely due to a number of high-profile security scares involving Microsoft products.
NVIDIA is now one of nine permanent board members, the other being 3Dlabs, ATI, Evans & Sutherland, Hewlett-Packard, IBM, Intel, and SGI. NVIDIA had been a term member. Apple, Dell Computer, Matrox, and Sun are term members. A term member still has a vote, but is on 1-year membership. This is a bit like the UN, as the permanent members vote (in closed sessions) on whom to admit as a permanent or term member. It comes as a bit of a surprise that NVIDIA wasn’t a permanent member before this. It also brings to mind “why now” questions. There was some contention between the ARB and NVIDIA over Cg and the GL Shading Language, with NVIDIA pushing an NVIDIA-centric Cg featureset and some others pushing anything but Cg.
Nonvoting participants include (as of June 2002) Alt. software, Crytek GmbH, Discreet, Empire Interactive, Ensemble Studios, Epic Games, GLSetup (which tells you this is an old list), id Software, Imagination Technologies (PowerVR), Intelligraphics, Micron, NEC, Obsession Development, Quantum3D, RAD Game Tools, Raven Software, S3/Diamond Multimedia, SiS, Spinor GmbH, Tungsten Graphics, University of Central Florida, Verant Interactive, and Xi Graphics. Microsoft quit the ARB to focus on DirectX issues.
|NVIDIA corp. acquired privately held wireless/mobile device graphics accelerator chip manufacturer MediaQ for about $US70 million. MediaQ, was launched in 1997 with backing from National Semiconductor, Weston Presidio Capital, Summit Accelerator Fund, Infineon Technologies, El Dorado Ventures and ViVentures. MediaQ sells its semiconductors and complementary software, API’s and drivers for major mobile operating systems (Microsoft PocketPC and SmartPhone, Palm, Symbian) to the main suppliers of mobile phones and PDAs including Mitsubishi, Siemens, DBTel, Dell, HP, Palm, Philips, Sharp, and Sony.
Not only has NVIDIA been using its muscle to move into the workstation market as of late, they now have the means to move into the hot mobile graphics market, apparently having something better to do than pine over letting the XBoxNext slip away. In all a very shrewd marketing move, particularly if they can sell 10′s or even 100′s millions of inexpensive graphics chips. Imagine if your phone or PDA had the power of a TNT2 driving its graphics – now that would be awesome! “As clever as Nvidia is, they probably couldn’t develop the same type of technology that MediaQ has in a reasonable time frame,” said Jon Peddie of Jon Peddie Research, Tiburon, Calif. “A company like Nvidia that makes big, powerful processors with millions of transistors does have a hard time scaling it down into something that’s small and doesn’t consume a lot of power.”
During a conference call with analysts NVIDIA’s president and chief executive Jen-Hsun Huang stated “The transition from 2G to 2.5G and 3G phones will drive the growth of high-resolution color displays and high-resolution cameras, creating demand for full-featured, low-power multimedia chips.” Huang estimated that of the 400 million to 500 million handsets expected to ship this year, the number with color displays could exceed 190 million, up from 54 million last year. Huang expects the acquisition to make NVIDIA a one-stop supplier of graphics and multimedia chips for the PC, consumer, and wireless-mobile markets.
XBitlabs reports some Mercury Research results for Q2 graphics market share from Q1. The big winner is Intel (even before they release the fricken Grantsdale chipset) increasing share from 27% to 32% – due to the integrated P4 chipsets. ATI increased from 20% to 21%, and NVIDIA fell from 31% to 27%. The report also said that NVIDIA had 60% of the DirectX 9 market share. Not bad for a company that botched its first DirectX 9 product release, although it seem that most of these were entry level (GeForceFX 5200) components. Still, it goes a long way to verify NVIDIA’s strategy of shipping only DirectX 9 capable products in its latest lineup. That leaves ATI with 40% of the DirectX market, though those are where most of the high end cards went. This roughly means 45% of the Q2 graphics market share was DirectX 9 capable cards.
The remaining 20% of the graphics market went mostly to SiS/XGI. Matrox Graphics, Trident Microsystems (now sold to XGI), S3 Graphics/VIA, Silicon Motion and 3Dlabs now occupy very small market shares.