It gets worse.

About a year afterwards, after having a truly fine dinner in San Francisco right around Christmas time, we were walking from the restaurant to the hotel and looking in the festive decorated shop windows, when this skinny guy comes up to me and says, “I’m not looking for a handout, but I’ll bet you I know where you got those shoes.” I was still enjoying the effects of an excellent Cabernet and said, “OK, where?” And, once again, I was five dollars poorer.

“Fool me once, shame on—shame on you. Fool me—you can’t get fooled again.” G.W. Bush, 2002.

Nvidia recently released a new driver, #180.84 by my last count, and with it improved their benchmark scores in certain games. ATI quickly reacted to that, lowered their prices and released their own benchmark results based on their most recent driver #8.12.

We love benchmarks as much as the next dweeb, and we have all the right gear to test things, and we have not only a web site to publicize the results, we have Tech Watch, which you may be reading right now, and we do it for free—no ads here. We get asked by the financial community and the press for our findings, we’re frequently quoted by those folks, and we’re in there with the testing stuff—we dig it and we get it Except that we can’t get it; consistency that is.

We can’t get no, con-sis-ten-cy

We can do a damn fine job with a repeatable, albeit synthetic benchmark like Vantage or 3DMark 06. And we’ve got a couple of game scenarios that we use (and are happy to disclose)—and that’s the rub—GAME TESTS ARE BUNK!

Back when Nvidia introduced the FX30 and it didn’t prove to be too good a part, Nvidia’s strategy was to blame the tool and succeeded in convincing the industry to test with real-world applications like the actual games we want to use. Fine idea, very similar to our own early benchmark tests for Windows back in the mid-eighties (called WITS for you history buffs). So we’re totally pro-game or application testing.

The problem is, there aren’t any. There’s nothing consistent or repeatable by anyone else. If Jerry from Timothy’s Benchmarking Fool web site calls 4D Cracker’s web site and says, Yo Xavier, how about sending me the script you used for testing Zombie Headblasters 101? Xavier will say, “kiss my pixels, do your own script you little maggot.” Remember, these guys are in competition to sell banners to the AIB suppliers—they’re not into sharing.

The solution

The game developers have to provide a test scene, as Crytek has done for FarCry and Techland has done for Call of Juarez that anyone can run with FRAPS or a built-in FPS counter. Until that happens you’ve got Balkanization of scores with each web site and AIB supplier using their own scene in a game to get test scores—what we have now.

A second choice would be for an independent to develop scripts for the games, much the way certain dedicated folks produce walkthroughs, and post them (and ask for a donation).

We need something that will allow true scientific testing—i.e., repeatability—in multiple labs.

The shoes

So what the hell does benchmarking have to with where I got my shoes? It’s a con my friends, plain and simple—make a monkey out of me—fool me twice.

Unless the exact same effects are being invoked every time, no in-game test is valid because it can’t be repeated, scrutinized, or evaluated for relevance. So the in-game scores are not scientific measurements—they are OPINIONS.

Epilog—Why game testing is bogus

Who is going to invest $200 to $500 or more for a graphics AIB just to play one game? If the graphics AIB from Blasphomere gets 65 fps in Stalker and Dismodyne only gets 55, but Blasphomere gets 45 in FarCry while Dismodyne gets 80, and you want to play Left Behinds 92, how will you choose? Wait for the Left Behinds 92 results before you buy a new AIB? Maybe, and then we see that Blasphomere got 25, and Dismodyne got 27. But all you want to do is play Left Behinds 92 and maybe later WOW 75. What’s going to give you a better sense of overall confidence, a repeatable (by you if you so desired) synthetic benchmark, or a totally unrepeatable game test score done by the Socko500 web page with low-low prices on all AIBs and power supplies?

Yet another solution—
a new rating system

How about the game developers or real gamer sites post a score on games that lists the loading on the GPU and the CPU? Then we could purchase a game on the basis of the system we have and/or make the purchase of equipment on the type of games we like to play.

The world is changing around us and anyone trying to live in the past with the rules of the past is doomed. The old sage about evolve or die is with us more these days than ever before, and it’s not going to get any easier.

Notebooks are rapidly replacing desktops, moving the venerable desktops into the specialty category of gamers and power users, which includes workstations and GP-compute systems—volume-wise the desktop is flat.

Handhelds have slowed in sales, and monitors are in steep decline, things are getting adjusted in the face of the new economics and the new investment strategies of major companies, consumers, and investment firms.

Therefore, anything you hear that may have sounded absurd, or ridiculous five months ago should now be given some consideration. One such idea that is floating today is an acquisition or merger of Apple and Nvidia.

With all the turmoil in the market and changes in fortunes between companies, Nvidia has shifted from being a potential suitor of AMD to being a possible consideration for Intel and, of late, a potential target for Apple. I can find some sense in an Apple deal. Unlike the rumors a couple of years ago when AMD bought ATI and some people thought Intel might buy Nvidia, an Apple acquisition is not ridiculous or impossible short of a few deaths.

The pros

Jobs and Huang have a mutual respect, they could (do) work together. Huang is a proud, smart, and energetic man and wouldn’t take to a bureaucratic and sluggish organization, and Apple is anything but that.

Apple has no clear heir apparent, so acquiring Nvidia and making Jen-Hsun CEO while ailing Steve Jobs moves upstairs to COTM would provide the environment Huang needs. Huang is ready for a bigger canvas and a combined Apple and Nvidia could be just the thing for his ambitions and vision.

If, as the Wall Street sharpshooters predict, Nvidia’s share price will never get above $10 again—and may have difficulty staying north of $5—Huang’s best shot at increasing shareholder value would be to accept an offer of $11 from Apple—no one wants to be the next Jerry Yang. With the market fluttering the way it is, striking a deal that would have to be renegotiated every 30 days would be the trickiest part.

The acquisition of Nvidia would swell Apple’s cash position giving it even great war chest for other acquisitions and/or R&D purchasing power.

Nvidia has the largest collection of the best graphics engineers in the world. Both Jobs and Huang respect engineering power and how it provides the platform for growth. And both Jobs and Huang have provided their engineers with the latest and best tools, so they are totally in synch on those points. Jobs would probably visibly salivate at the prospect of Nvidia’s engineers being part of Apple.

Apple clearly sees the value and power of the GPU and can easily imagine offering computers with Atom processors and powerful graphics. Scenario two of this premise is that Apple also buys VIA and gets total independence. A kind of back-to-the-future of vertical integration, something Apple has shown a desire to do in many ways from manufacturing to marketing.

A GPU product line would fill out Apple’s self sufficiency and expand it into newer markets with the kind of margins it desires—and its shareholders demand—in order to sustain the high PE that it has.

Nvidia’s Tegra becomes iPhone III and/or the game machine I’ve speculated Apple will bring out. And, as good as it is, Nvidia’s UI can teach Apple something about UIs—the Windows OS may be a small problem, but easily overcome.

To round things out, Apple launches a line of AMD Shanghai-based HPC servers with Telsa GPU-compute racks using a combo of OpenCL and CUDA.

Why not?

Jobs can’t/won’t share the spotlight with anyone.

Apple offers an insulting (sub $10) price.

Even Apple isn’t a big enough environment for Huang’s vision and ambitions.

Conclusion

Apple is in the best shape it has ever been in. Nvidia is bruised but still an extremely valuable company with strong resources in IP, people, and cash and would be an amazing asset for any company assuming a suitor would be acceptable to Huang and the Nvidia BOD.

Apple and Nvidia need to get out of the shadow of Intel to thrive. Apple is a product company that needs components to maintain its vision and independence, and the components have to be world-class with long legs—what company beside Nvidia could claim such a position? The combo would ignite Apple’s share price and stabilize speculation about the future for Jobs the thaumaturgy. And Huang could take Apple to new heights.

I like both companies a lot, I really like the notion of the synergy between them.

If you add up all the shipments of the recent big recent-generation consoles (includes PS3, Wii, and Xbox360, excludes DS and PSP, PS2, and all others) you get something close to 74.7 million units. That seems like a lot.

The PS3 and Wii came out in November 2006, and the Xbox 360 was introduced in May 2005, so the 74.7 million is over several years.

From Q3 2005 till Q3 2008, 196 million gaming PCs shipped, and 10 million of them were in the Mainstream Enthusiast class—and if we include mainstream desktop and High-end high-end notebook, we could add another 187 million units—now that really is a lot.

In our latest series of reports, The PC Gaming Market, where we talk about the three classes of PC gaming machines and their market, what we found was that it’s a huge market. And the tie-ratio for PC gamers is higher than consoles.

So I get a little tired of hearing about what a superior market the three-to-four-year-old consoles are when more PC gaming machines with the latest technologies are being sold every day.

Figure 1: PC Gaming market worldwide in billions of $US dollars.
(Source: Jon Peddie Research PC Game Market report)

Now, the question is, why doesn’t PC gaming get more respect? Why isn’t it recognized as the viable, growing, and huge market it is?

Part of the problem is that PC gaming has been associated with first-person shooters, and the PC gamer has been depicted as some fat white boy with zits and an enlarged right hand wrist muscle from either mousing or personal dating. Whereas the console gamer is depicted as a family member who may zap a few bad guys in Halo, but is really more interested in dancing with her mom and dad in front of the Wii—image is everything.

Also, the game developers, the ISVs, and distributors prefer to write a game once and have it play for three to five years on a stable (i.e., dead-end) piece of hardware. That’s how they get the most ROI.

Developing a new game for a PC, even if it is just another sequel to the last game, is hard work. And who really cares about that cinematic-quality crap anyway, you’re just going to look at it on a TV set right? What can you see on that?

Thankfully HD TVs are sneaking into the living room bringing 720p and, in some cases, 1080i capability. Imagine, 720p almost as good as 1024 x 768—old XGA resolution introduced by IBM in 1988—wow.

The average PC gamer has a 17-inch screen with at least 1280 x 1024, most serious gamers have 21-inch screens with 1680 x 1050, and a few of us extreme players have 30-inch screens with 2580 x 1600—almost 2x the best 1080p HD TV and with comparable screen size.

Yeah but, the lazy developers yell, a console sells for a few hundred dollars, and a PC sells for a couple of thousand.

Horsefeathers. Let’s compare. There’s two ways to do it. We could do 2008 to 2005 dollar comparisons (i.e., what would $250 2008 dollars buy in a 2005 PC), or we can do it on a MIPS/FLOPS/screen res comparison. And, when you’re not using your PC for gaming, you can watch a DVD just like you can with an Xbox 360 or PS3 but, oh, sorry, not a Wii. Or you can do real work, not just some feeble web browsing in a closed garden.

Fact is the PC Gaming Market is bigger, worth more money, growing faster, and has better technology than the console market. How come no one but us seems to know that?

Apple began in 1976 with the 6502 processor and, in 1984, moved to the 68000 with the introduction of the Mac. The company introduced the new computer with an amazing video that played off 1984—that commercial cast the tone and course for the company for the next 34 years. In 1994, Apple partnered with IBM on the PowerPC CPUs and stayed with IBM till 2005 when Steve Jobs officially announced that Apple would begin producing Intel-based Macintosh computers beginning in 2006.

During the last four or so years of the Apple/IBM relationship, Apple kept pressing IBM for higher performance CPUs. Apple needed to more effectively compete with Intel-based PCs. After too many broken promises, Steve Jobs, who had returned to Apple in 1997, had enough and made the switch to Intel. Jobs takes quality and performance very seriously, his customers know it, and his suppliers know it—and they don’t get to stay suppliers very long if they don’t deliver.

With that as a backdrop, I pondered the recent move by Apple from Intel’s IGP chipset to Nvidia’s. Could it be that Apple was tired of waiting for Intel to catch up with ATI and Nvidia? The G45 is a prime example. It has two cores, whereas the Nvidia part has, er, 16!

And the Intel chipset is that, a set, i.e., two chips, a northbridge/IGP and a southbridge, while Nvidia’s chip is, well a chip.

So I was wondering if the irony of this event is that the company that Apple dumped IBM for is also being partially dumped. In Appleland, you either put out, or get out.

On battery life

One of the best laptops I ever had was a 1993 HP Omnibook 300. It had a 9-inch passive B&W (actually light and dark grey) 640 x 480 screen, and a built-in absolute position mouse. It was equipped with an all solid-state memory (10MB flash), the OS (Windows) was in ROM and it booted in a few seconds. It held state and came to life in a second and its 386 processor could run all Office apps. It would work flawlessly on a transcontinental flight and I think even on a transatlantic flight from SFO. Wonderful machine, weighed almost 3 pounds. I used to carry an extra battery and never used it. In a pinch, the Omnibook 300 could also run on four AA batteries. It cost $1,515—the equivalent of $2,150 today.

Nine hours on a charge, held state, OS in ROM—15 years ago. Maybe we’ll get back to that future one day.

I use the Omnibook 300 as a yardstick when evaluating new laptops. The HP Mini-Note comes close.

In those days and, in fact, up until about 2003-2004, the criteria for a laptop was to run on batteries on a transcontinental flight—i.e., four to five hours. And, it was assumed you’d be doing spreadsheets, word-processing, and maybe email during that time, and possibly view a presentation. Then, somewhere along the way, the idea was introduced that you could also watch a movie on your laptop. At first the reaction was, why in the hell would I want to do that? But with the introduction of HD+ 15-inch screens and decent GPUs, CPUs, DVD drives, and pretty good sound, it became a reality—and popular. I can still remember how surprised I was to see someone on a plane actually doing it—and his seat mates couldn’t help looking over at the screen.

Reality check

The average feature-length movie is about two hours. Two hours—not five, not nine. And IGPs have gotten good enough that you can play a movie using them. So what’s the big freakin’ deal about “able to run a feature length movie on a single charge?” Two hours.

Talk about reality distortion—since when did being able to use your laptop for two hours become better than running it for four or five hours—what am I missing here?

In 1999, the Internet bubble began. Most people knew it was a bubble, laughed at the silliness of most of the deals, but a greed factor and feeding frenzy, fear of missing the boat and being left outside took over logic. Three years later the stock market crashed (relative to the run up in 1999) and didn’t get back to the pre-bubble level until near the end of 2003.

Today, the market is below the pre-bubble mark (9,300) effectively wiping out 10 years of growth, despite the tumultuous swings in between.

Share prices are based on expectation, and so they typically carry a value greater than the net-worth of the company. The difference between the value of the stock and the stock’s price is a function of the price-earnings ratio, or PE. A high PE means the investors expect great things for the company down the road. How far that road is varies from investor to investor. For some it’s 10 years, others five—it’s not an exact science.

If we look at IBM, one of the best run, most profitable companies in the world, we see its share price has more or less tracked the DJIA.

Last week, IBM said that it earned $2.8 billion, or $2.05 a share, on revenue of $25.3 billion, up from a profit of $2.36 billion, or $1.68 a share, on $24.1 billion in sales during the same period a year ago.

Analysts had forecast IBM to earn $2.01 a share on revenue of $26.5 billion. While IBM’s earnings topped analysts’ forecasts, sales fell short of the last consensus estimate, but several analysts were said to be in the middle of adjusting their revenue targets lower when IBM issued its earlier-than-expected
results.

IBM also said that its gross margins for the quarter rose to 43.3% from 41.3% a year ago.

Now, that’s pretty amazing, and yet IBM’s share price dropped from $103 on 06 Oct 2008 to $91.66 on 10 Oct 2008—a change of -11%. So even though IBM did really well this quarter, the investors think IBM’s prospects for the future (that five to 10-year road) aren’t so good. You’ll notice in Figure 1 that IBM got hammered at the Internet-bubble crash in 2002 too, and then came back to pre-bubble values very quickly, so back then, some investors must have decided they were wrong about IBM and changed their bet.

Figure 1: The DJIA reflects the mood swings of investors, but has little to do with reality.
(Source: Jon Peddie Research)

Figure 2: DJIA vs. IBM.
(Source: Jon Peddie Research)

So IBM, and others, who had nothing to do with the deregulation, the lack of assets in AIG, the dishonest brokers and loan offices, and the stupid panic driven home buyers, is having its stock price devalued because investors fear it won’t be able to borrow money to operate. IBM has $9.8 billion in cash reserves—that’s more than 124 of the 224 countries of the world have (e.g., Bolivia is at $9.2 billion) so how much borrowing does IBM have to do?

True IBM’s sales, like others, may suffer a slowness in growth, maybe even a flattening as consumer confidence wanes from all the bad news and the machinations of the U.S. congress and other countries’ national banks.

The world financial crisis now ranks with the 1987 stock market crash and the 2001 dotcom collapse in its impact on consumers, pushing their confidence levels sharply down this month. The U.S. consumer confidence index dropped 32 points in October—its largest single-month drop since the Consumer Attitudes and Spending by Household Index began in 2002. The index is at 37, compared with 69.2 in September. And this is going to have an impact on retailers as a result of the domino effect of the mortgage crisis, rising energy and food prices, higher unemployment, and now the investment and banking crisis. The lack of consumer confidence has eroded to a point where people are afraid to spend any money on anything and that will ripple back to the computer, IT, and other big industries.

Are computers to blame?

But so much trading is done automatically, based on complex and almost unfathomable algorithms that I have to wonder if the singularity has already happened and the machines are running the world.

Hedge fund managers, day traders, institutional investors, and retirement fund managers seldom make buy-sell orders anymore, or even the decision. They check the screen, sniff the air and press Enter. Then, within microseconds, a trade is made.

How are we going to get out of this mess if the computers are in charge? Where are the folks who wrote those programs? Do the hedge fund managers and investment bankers even know? Are we now at the mercy of a bunch of nerds? Can anyone see a way out of this mess?

Or is it over? Can’t we re-boot?

3D for entertainment, 3D for signage, 3D for science and engineering, 3D for defense, 3D forever. We, happy consumers who will do just about anything we’re told, will now have a new fashion accessory—3D viewing glasses. We will have multiples of them, different sizes, shapes, and colors to fit the clothes we are wearing, or maybe not wearing at all considering some of the uses of 3D.

3D, of course, is not new, and to borrow a page from Kathleen’s book, it has sputtered and stopped and sputtered and stopped again due to the inability of its promoters to cross the practicality gap.

But today’s promoters and supporters tell us, yell at us, that that’s all solved now, all those problems have gone away, been resolved, and it is the new wave, the new thing, our lives will never be the same. Sound familiar? I do believe we have been promised such things before.

The powerful people cite the inevitability of this new wave. They say that just as talkies replaced silent film, and color replaced black and white, and multichannel sound replaced mono, 3D will be the way we watch all, not some, but all movies in the future, and that future is rushing towards as giant studios like Disney and DreamWorks commit to 3D 100%.

Right behind them, or maybe in synch with them, are games, PC games to be exact. PC games, modern PC games, come with a built-in advantage, they are already constructed in 3D, so getting them to display 3D vision is easier, or so we are told—there is that little issue of refresh rate and 60 Hz LCD displays, but not to worry, that will all be taken care of.

3D TV will have to wait a while as standards and displays get developed. 3D TV is theoretically possible now with high-speed DLP TVs and projectors, and high-bandwidth media like Blu-ray. So, we can don our evening wear 3D glasses, pop in the season collection of Battlestar Galactica, and bask in the joy of missiles, Vipers, and number six protruding into our living room. Once number six is projected in 3D there will be no going back.

Well, I’m still waiting for the share price of my dot-com companies to come back up to the purchase price so I can sell them and put the money back in my retirement fund. I’m not really in the market right now for the newest old best great thing. I’m not convinced the practicality gap has been crossed. One of the reasons I’m taking that position is the abundance of solutions. There are five choices of glasses. That’s three too many. We can have two types, one for movies and one for PC and TV, but not five. Before 3D is the defacto entertainment standard, it first has to get some—standards that is. It also has to overcome the chicken-egg thing of installed base of the necessary technology vs. the content. And then there’s that content issue. One of the things that killed, well I guess it didn’t kill it, but at least severely wounded 3D back in the 1950s was the awful content. And, with the exception of some of the IMAX stuff, it has remained awful, poking things out at the audience, no story, and then the ultimate eye fatigue and disorientation. One of the most recent movies the studios are raving about, Journey to the Center of the Earth, is so bad it’s amazing it got released.

Sports, on the other hand, do look good, better, and more interesting in HD and 3D. Sports, for the most part, aren’t staged, and there is no director yelling throw the ball at the camera. And sports, especially fast-paced sports like basketball and soccer (football) are especially challenging to the refresh issues, but, when they get it right, they are truly thrilling to watch.

So that’s my prediction. 3D will not become a mass consumer home entertainment vehicle for quite some time, but special-event 3D will. Special sporting events, special well-made movies, and maybe even a well-made computer game will be the market.

So don’t rush out and buy your designer 3D glasses just yet, be happy with the ones the theater loans to you for a couple of hours, which ever type they may be.

Epilog—movie director that I want to be, I have observed that the 3D effect in movies is more believable and more effective when things point into the screen, away from the viewer, rather than out and at the viewer.

We now have four offerings for the consumer space of multi-core chips, which are euphemistically called processors. That is just a taste of how confusing this is all going to be, especially when the market spin is added. For example, AMD’s four core chip called Barcelona is called A processor, as is Intel’s i7 Nehalem. But those “Cores” are in fact full-function multi-stage out-of-order 64-bit double-precision floating-point processors. So the term “processor” has come to mean chip, and the tem core has come to mean processor…a rose by any other name. It doesn’t really matter what we call these things as long as we all call them the same thing and share a common understanding of their definition.

Sticking close to common usage and not inventing or adding to the litany of words for the lexicon will be an underlying goal, if not rule, for this discussion.

Then, we have GPUs with hundreds of 32-bit double-precision floatingpoint processors, but they’re not called cores. They’re called streaming processors and sometimes SIMD arrays. And the competitors of GPU suppliers call the GPUs, application-specific integrated circuits (ASICs) or “accelerators.”

The GPU suppliers openly admit that within their GPUs there are indeed some accelerators, most notably video-processing accelerators for MPEG. There are also texture processors.

Intel, as you know, is coming out with a multi-core, or a many-core processor called Larrabee, which will have between 16 and 64 cores (depending upon SKU). Since the cores in Larrabee are x86 units with a floating-point processor, they are called “cores,” by this new unwritten law we’re developing. However, Larrabee also has hardware accelerators in it for texture processing, so does that make it an ASIC or an accelerator? Not if Intel’s marketing folks have anything to say about it—and they will.

And so we enter the newest battle of words, the newest battle for mindshare and, of course, the all-important battle for market share.

I made a comment a couple of months ago that it was ridiculous to compare Larrabee to a GPU. I was quickly challenged on that remark by an Intel person and asked how I could make such an uninformed statement. In public, no less? My response was how can you compare a 240 processor Nvidia GPU or an 800 processor ATI GPU to a 16- to 64-core Larrabee. There’s no comparison—the GPUs have between 40 to 60 times as many processors. My challenger smirked, somewhat snidely, and said, “Well, I guess it depends on the quality of the core doesn’t it?” I pondered that but never resolved it.

Later, I was speaking to an Nvidia person and said ATI’s new GPU will have better performance. The Nvidiette quickly snapped and said, “What do you base that on?” Well, I said, for one thing it has 800 processors and you only have 240. “Well,” said the Nvidiette, “I guess it depends on the quality of the core doesn’t it?” I thought, there’s an echo in my head, I should see a doctor.

So, now not only do we not have a common term for a processor, or a core, or an accelerator, but along with it, we seem to have a quality issue to contend with.

Of course these are just the opening barrages of the three giants, trying to confuse and affiliate the masses to their cause. That confusion, or spin, will have some effect, because all of the consumers will be the blind men following the one-eyed chip makers. And a bunch of them will buy stuff from all three. And in time, perhaps it will take a year, the logical and correct winner will emerge; the company that is providing the best performance for the best price with the least heat and space.

All three will tell you it’s them, and it’s now, or very soon, trust me.

So, we will have to swim in these choppy waters looking for a life raft and not knowing what to call it when we find it.

Having been dazzled by four days of dancing delightful pixels in LA at Siggraph, I then took a shower and went to San Francisco to continue the consumption of pixels and information at IDF. Another shower and then off to Palo Alto for pixels and cores, lots of both (burp). Skipping the shower, I rushed to join the rest of the JPR team at the inaugural Nvision conference in San Jose—a three-week smorgasbord of west coast pixelarity.

Pixels, we’re immersed in them about 17-7-365. From the really big cinema screen, to the living room HDTV screen and the work/play computer screen, down to the tiny handheld screen that is either a phone, the display of a digital camera, a portable DVD player, TV, or a GPS device. In addition to all those personal pixel blasts on our overworked cornea, we are visually abused by digital signage in airports, shopping centers, Las Vegas city streets, and gasoline pumps. So many pixels, SO many pixels.

How many?

Let’s start with us. How many pixels can we process? It’s not an exact number, but if we assume our general response time is 30Hz and that there are 120 million rods and six-million cones in each eye, then that’s 252 million potential pixels sampling 30 times a second, which give us a 7.6 billion pixels/sec processing capacity.

All those pixels get pumped through our 12 paired cranial optical nerves where the impulses get dumped into the visual cortex in the occipital lobe of the brain. Just as a reference point, if you had a display that filled your visual field of view (assume a 30-inch monitor in a dark room) then you’d be getting fed 124 million pixels a second, so we’ve got capacity to spare—or do we?

Every year, we build and ship zillions of pixel-generating devices from LED clusters, to plasma tubes, LCDs, and CRTs. Did you ever stop to think how many pixels actually ship a year? Well it’s difficult to get a total total count, but I took a shot at it.

Six-hundred twenty-six trillion pixels shipped in 2007! And that’s probably conservative. My estimate is there are over five quadrillion (5x1015) pixels being pumped out every day.

Pixels? Bring em on

And if we assume half the world’s population of 6.6 billion people are seeing these pixels with their 7.6 billion pixel-per-second processing bandwidth, then all of them combined represent a pixel-processing capacity of 25 quintillionpixels per second, (25 x1018) and that leaves us with more than enough capacity to soak up all the TV, signage, PC, and other pixels being blasted at us every waking minute of everyday.

So bring it on display industry, give us your best shot.

Now, if only 1% of those pixels had any real value or meaning, then we’d have to calculate our ability to consume information, thank goodness for soaps and popups.

You’ve all heard me say, “The more you can see, the more you can do.” It’s Peddie’s Second Law. I have four screens on my desk; I think you’ve all seen the pictures. There’s one to left of my laptop, one to right, and above it. The screens are 20-inch 1680 x 1050. I use a track ball to mouse around (doctor’s orders). And I spend too much time spinning that little ball getting from the far left screen to far right or elsewhere.

My laptop is an HP dv9050 and sitting in the center top bezel is a camera. I’ve used it on occasion to chat with Kathleen over IM. Other than that, it just sits there quietly watching me like the red eyed HAL.

I want HAL to work more, earn his keep. I want HAL to follow my eyes and, when I look away from the document in focus (i.e., selected), I want the mouse pointer to go where I’m looking. I’ve got Smart mouse turned on so it snaps to the default button in a dialog box when one comes up, and that helps. It would be better if all programs, including Microsoft’s, worked the same way so the mouse knew when an action came up, but that’s a rant for another time.

Now, such an operation will be tricky, but I think doable. As long as you don’t change focus on an app, the mouse pointer can go anywhere without affect or effect. So if, in the course of creating a document, my eyes wander (as does my attention) there’s no harm done by the mouse arrow dancing around on the screen. And, as many of you know, you can have more than one mouse in operation. You’ve learned this when your thumb rests on the touch pad of your laptop. The protocol is that the last mouse wins. In my eye tracker, manual mouse always wins, so there needs to be a touch sensor on the base of the mouse to indicate that you’ve got your hand on it. The operation would then be like this. I’m loading numbers into a spreadsheet which is opened on the right-hand screen. I need to copy a number from a web page that is on the left-hand screen. I look at the left-hand screen and find the number, and the mouse pointer following my eyes gets near it. It doesn’t have to be exact. I lay my hand on the mouse, and the eye tracker releases the pointer leaving it where it is, and then I use the mouse for fine control. I now have the copied number and look at my spreadsheet (with my hand off the mouse) and snap-a-rudy, the mouse pointer pops over to the spreadsheet, etc.

Now if a $200 Wii can understand what my hands are doing, and a PSP, with the help of a cheap camera, can see my head up close and personal, my web cam should be able to see my eyes and follow their direction. It’s easy. There’s that little blue dot in the center of that big (mostly) white glob, and it’s redundant, one on each side of my nose. And a touch sensor in a mouse is nothing, a simple capacitor.

So get to work Logitech, Microsoft, and HP. Help me do more by seeing more and mousing less.

Epilog

Some of the technology I’m looking for already exists, and some of it is being used in physiological studies.

• Eyetracking Research
• Sony 98047 Playstation Eye Ps3
• Wikipedia Eye tracking
• EyeTracking Inc. Technology with a vision

While waiting for someone to build my eye tracker, we’ve experimented with a head tracker that requires me to turn my head, and learn to be more precise in my positioning. I don’t think this is the final answer.

Adobe and Vista rant

A Dumb Obtuse Balky Executable, ADOBE, crashes a Very Incompatible Stubborn Touchy Application—VISTA. Speaking of mousing around, I have a problem with Adobe 9 and Vista 64. Clicking on a link that requires Acrobat to load often causes my system to freeze, and I lose my mouse. What good is a computer without a mouse? I lose the keyboard too, so it’s a dead machine with all the displays lit up and nowhere to go. I know it’s my fault. After all, I’m asking two alien entities to work together at the same time in the same space—how dumb is that?

But I’m trainable. I have now learned to postpone any calling of Acrobat until I’m finished doing anything important, make sure everything is saved, and then try CALLING Acrobat. If it crashes (about every fourth time, randomly) I can recover. But my doctor says I’ve got to cut down on the coffee. “Find something else to do when your machine crashes,” he said.

The mechanical loom developed by Joseph Marie Jacquard between 1801 and 1805, which used holes punched in pasteboard, was the beginning of the digital age. Were it not for that machine, we’d still be calculating using gears and analog representations. One can only imagine (and what a fun mind experiment it is) the kinds of automated writing machines we might have if we were stuck in analog land.

The move to digital, first mechanically then electro-mechanically with solenoid and relays, then early electronically with vacuum tubes (valves); GTX280 the biggest development since the industrial age. Or, more accurately, it is the consummation of a process that began in the Industrial Age and is hurtling us beyond the Space Age. Digital has empowered the internet, computer graphics, and uncountable other devices and industries.

And whereas almost everyone praises the Internet as the greatest development in the world since the stream engine, in fact it was only possible due to the advent of digital mechanisms and communications.

From Jacquard’s loom in England it took 130 years to get the first digital, or binary, computer developed by Konrad Zuse in Germany, in 1938. Prior to Zuse, mechanical calculating machines of that time were based on the decimal number system, including Babbage’s amazing machines. Zuse’s machine used no relays but rather a novel mechanical memory device consisting of thin metal strips.

Things moved rapidly after that and the first electronic binary computer was built at Iowa University of Iowa in 1937. It was a hard-wired machine and not programmable, and it’s highly unlikely John Atanasoff's and Clifford Berry had any idea what Zuse was doing due to limited communications and political difficulties of the times.

Figure 1: The Z1 computer with manual crank for driving the clock byhand.

The Colossus, the first programmable, binary electronic computer was de­veloped in the UK at the famous Bletchley Park in 1943, where it was employed to break German codes, and from it came forthe several improvements in England and the U.S., and ultimately all over the world.

Jacquard’s loom had a few dozen moving parts, made of wood. Zuse’s Z1 had 30,000 mechanical parts, and the Colossus had 1,500 valves (tubes) plus a few hundred relays and various other electro-mechanical parts for programming and printing.

Compare that to today’s GPUs from Nvidia and AMD. Nvidia’s GPU has over a billion transistors, AMD’s close to a billion. It gives you a sense of how far we have progressed in 65 years. And in 10 years, maybe a few more, we’ll talk about the zillions of quantum states in our current machines as we approach the singularity.

The acceleration of technology and all the benefits and, in some cases, heartaches it has brought is completely due to the adoption of the digital binary system. Although never postulated as a seminal discovery like electricity or e=mc², we could not be where we are without it. The colossal irony of it is that we human animals are decimal and analog based, and we have had to abandon our natural preference for such domains in order to exploit the power of digital binary mechanisms—clearly one of the great demonstrations of human adaptivity. I think there is hope for us after all.

For if you’d like to read the fascinating story of Konrad Zuse, you can find it at http://www.epemag.com/zuse/default.htm#index.

Boy, for as long as I’ve known the company, which is actually before it was a company, I have never seen such a torrent of bad news and bad breaks at one time. Here’s the list of recent events and news:

• Gave a guidance down from a previous forecast.
• ATI’s new products offer better price-performance challenging Nvidia’s price platform.
• Price cutting on the new and older parts.
• Set aside a reserve for parts failing due to temperature cycling.
• Delayed ramp of a next generation MCP.
• The company is being sued by Rambus.
• Intel’s continue war of words declaring the GPU as archaic architecture.
• Share price dropped 60% (normalized to the DJIA) since the first of January.

The quote from the press release reads “The estimated decrease in revenue and gross margin is due to several reasons: end-market weakness around the world, the delayed ramp of a next generation MCP, and price adjustments of our GPU products to respond to competitive products.” Ouch.

Downward guidance

The trouble began when the Company had to announce that second quarter revenue and gross margin are expected to be lower than guidance provided during its first quarter financial conference call and be from $875 million to $950 million.

The company had previously forecast that revenue would be down 5% sequentially from the April quarter, or about $1.09 billion. The company previously said it expected gross margin to be up about 100 basis points from the April quarter. Nvidia said the shortfall is due to end-market weakness around the world, the delayed ramp of a next-generation media and communications processor, and price adjustments to its graphics processors to respond to competitive pressures.

The second quarter is always down, but last year Nvidia defied gravity, and may have gotten over confident that they could do it again this year—however, they probably didn’t count on a looming recession.

ATI’s new products

ATI surprised everyone and came out with high-performing parts in the midrange and the high-end at very aggressive prices.

Nvidia tried to counter in the midrange with a new 55nm based part, the 9800GTX+ initially priced at $230. However, that had the effect of impacting inventory of 9800GTX products in channel at $400 to $500, with the over-clocked versions being $500 and equivalent in performance to the GF9800GTX+. So, with the overstuffed channel (lots of 9800GTXs available out there) this looks like a train wreck because there’s price protection for the partners, so all this is going to impact margins for a couple of quarters.

Price cutting

Nvidia also had to drop prices on its new GTX2x0 AIBs to meet ATI’s aggressive new prices. The AIB (board partner) prices are expected to be as follows (US$)

Set aside

With regard to the one-time charge from $150 million to $200 million against cost of revenue for the second quarter to cover anticipated warranty, repair, return, replacement and other costs and expenses, arising from a “weak” die/packaging material set in certain versions of its previous generation graphics chips and chipset products used in notebook systems, well that could happen to anyone. The webzines are having a field day with this, some claiming Nvidia tried a cover up and that all their G84 and G86 parts, notebook and desktop, are bad. Nvidia says certain notebook configurations with chips manufactured with a certain die/packaging material set are failing in the field at higher than normal rates. The company plans to seek insurance coverage for this matter. The web­zines claim it’s mostly HP notebooks, and suggest some Dell units may also be involved.

The costs to Nvidia will be offset to the OEMs for their service costs to replace failed units, and to push a new driver that keeps the fan on longer (to minimized the temperature cycles—but which will also lower battery life). Some wags have suggested, and we can’t help but agree, that this is a class action waiting to happen, and that Nvidia may need a bigger set aside.

“This has been a challenging experience for us,’’ Said Jen-Hsun Huang, Nvidia’s unflappable president. “As for the present, we have switched production to a more robust die/package material set and are working proactively with our OEM partners to develop system management software that will provide better thermal management to the GPU.”

Delayed ramp: As for the MCP missed production cycle, it was/is for the Intel Penryn CPU with FSB. No, it’s not 790i that’s been delayed—its the discrete chipset. The delay affected the motherboard GPU, which the company hasn’t announced yet, so I won’t steal their thunder. Expect it ~August though.

Rambus suit: And finally in what seemed like a piling on, Rambus took the moment to announce that they had filed a suite against Nvidia over alleged memory controller patent infringements. Rambus alleges six Nvidia product lines infringe 17 patents it holds covering various graphics and multimedia processors in the suit, filed in the U.S. District Court for the Northern District of California.

Intel’s war of words: One of the bright spots in Nvidia’s future is GPU-compute, and the company is investing heavily in its CUDA programming environment. However, in a swipe at that, Intel’s Sr. VP Pat Gelsinger, relegated CUDA, and the Cell, to a footnote in the history books. Gelsinger has said that it sees no place in the future of computing for general purpose GPU (GPGPU) programming models such as Nvidia’s CUDA. That, of course, was not taken lightly by the investor community who freak out anytime Intel mentions Nvidia’s name or products.

Share price: As a result of all of the above, Nvidia’s share price took a hit that knocked 30% of its value down, putting stock options under water and scaring the hell out of 401k managers. The result has brought Nvidia’s year-to-date decline to 60%. The stock closed at $11.67 Friday, 11 July, at a two-year low.

Table 1: Nvidia’s share price this year—down 66%.
(Source: Company data)

When a company has been such a stellar performer for so many years, it is almost as if some folks got some kind of giddy enjoyment out of the rash of woes that suddenly hit the company. Certainly the Inquirer has taken great pleasure in announcing Nvidia’s demise.

The move to 55nm: Nvidia taped out the 55nm GT200 a few months back and we believe that they will pull this product in as fast as they can to offset costs and performance gains by ATI. It’s unlikely they can get it in time for back-to-school, but the holiday season is a possibility.

I was asked recently what do gamers look for in a board, and I came up six items:

• Performance per Pixel (dots)—Max res and max performance.
• Performance per Price (dollars)—Max performance per dollar.
• Performance per Pressure (decibels) —Max performance quietly.
• Performance per Package (dimen­sions)—Max performance in a small form factor
• Performance per Pins (doubling)—Crossfire/SLI.
• Performance per Power (Watts).

Table 1: P3 scores for recently introduced AIBs.
(Source: Jon Peddie Research)

Figure 1: Compound percentage difference of three parameters: Price, Performance, and power—P3.
(Source: Jon Peddie Research)

Not all those parameters can be easily measured or obtained, but three of them can: Performance per pixel, price, and power. But which one is the most important, because they are all in different units of measurement and will have to be scaled? I decided that’s the individual’s choice and so made up a model with them weighted equally.

I generated a table of values, applied a weight to them, multiplied all of them and came up with a score.

Then I normalized the data to the lowest score and generated a chart. The performance data came from our testing using Futuremark’s Vantage and running on the Intel Skulltrail PC.

There was a great temptation to call this the PeddieMark, and play more on the Ps in the parameters, but as you can see I’ve resisted doing that. I’d much rather a useful metric actually used than my name in lights.

“Great jumpin’ Josaphat, have you learned nothing in the last 10 years,” I yelled at him whilst skillfully avoiding a drunk in the middle of the street yelling Jesus loves me, “Of freaking course we need it, haven’t you heard of suspension of disbelief, like, dUH!”

“Yes, Jon,” he said in that Catholic-school-boy-patience-is-a-virtue-tone he gets, “I know all about freakin’ suspension, but you’re not listening.” (At least he didn’t say, “as usual.”) “Why,” he said, overriding my attempts to interrupt him with my professorial lecture, “in the hell would we want it to be real?” Again I tried, unsuccessfully to yell logic at him, “Why,” he continued, “do you want to make it believable that you can go around and blow the heads off people?” Had a fly been in the car it surely would have flow into my open mouth which was struggling to form the word, because…

“Ah,” was the best I could do, “Ummm,” and it got quiet in the car except for the background mumble of NPR.

We drove like that for a while, not far because it was rush hour traffic, but several years in a fruit fly’s life.

And then the brilliance of his catholic school education clicked in, and as my friend Frank says, the nickel dropped.

Now, in addition to dodging the crappy drivers in San Francisco’s rush hour, I had to quickly find an argument to prove that this was my idea all along. It wasn’t working. And suddenly my periphery grew dark as the world around me closed in. “OMG, you’re right. We don’t want GTA4 real. OMG. What the hell were we thinking of?”

“You want realism,” the grasshopper said instructing the master, “when you’re playing large games like Stalker maybe, definitely Hockey, all kinds of sports. You want suspension of disbelief when you’re flying Ted’s Blackbird, or racing a Formula 1, but you don’t want it when it’s blatantly anti-social, when it goes against everything we’ve been taught in a Judeo-Christian society.”

I nodded my head struggling to produce an intelligent contribution to his thesis.

“But, he continued, oblivious of me, the traffic, my fine automobile, and everything else, “most of all you want story. And when you have story, really good story, you don’t need special effects. Isn’t that what Kathleen says?” Great I thought, now they’re ganging up on me, and she’s not even here.

“Ah,” I managed to get out.

“Look at Dagon’s Lair, back when we played games on arcade machines, back in 1983” he continued. “Because of the involvement of Don Bluth, a former Disney animator, the artwork in Dragon’s Lair was the highest quality ever seen in an arcade game. I don’t think it’s been matched. You don’t need special effects, global lighting, ray traced hair, soft particles, and water effects if you’ve got story. Story is what suspends disbelief. Did you have special effects when you read science fiction as a kid?”

Clearly the student had become the teacher, Aristotle would have been proud of me, but I was too consumed with avoiding sofas that had fallen off trucks, and trying to find a crack in his argument to give a good damn about some long dead Greek.

But his logic was inescapable. Some things demand and will never be right without absolute and ultimate realism, and others, realism not only distracts from the creativeness in generating story, as Kathleen complains about so many new movies, but is even more sinister in terms of the sociopathic attitude it creates in young plastic minds. Not just young minds, all minds.

When surveyed about the carnage in violent movies, back in the day when kids were allowed to watch them, the kids blew it all off and said, “Duh-ah, it’s a MOVIE, y’know.”

And the game publishers of billion dollar franchises like Rockstar/EA say the same thing—“Duh, it’s just a game—the kids know that.” I guess we can only hope they do. My grand kids seem to get it, but still their mother and I worry.

But Robert’s right. I don’t like the gore in some FPSs, and although I don’t turn it off, I secretly worry about the teenager down the block looking at it and thinking it’s OK to blast someone with a shotgun.

Realism: it may not be the Holy Grail after all.

For a while there was peace in the valley. The PC market had gone through its consolidation, we got over the internet bubble, the platform was stable, everyone knew their place and there was a place for everyone. Life was good, margins were good, growth was good, PEs were good, people smiled at each at each other at conferences, and good cheer was shared by all.

And then things changed.

Intel decided AMD shouldn’t own the power crown (performance or watts), or the 64-bit crown, or the internal communications crown, and those pesky GPU things were growing ASPs and market share too fast—this had to stop.

The GPU suppliers decided their zillions of floating-point processors could out process the puny single floating-point processor in a scalar X86 and had the audacity to tell people that.

And AMD, fresh from its acquisition of ATI didn’t think CPU integration should stop with the memory manager and announced that the GPU would soon join it.

Intel countered by announcing it too would have internal graphics, and what’s more, a super coprocessor named Larrabee, so there.

The GPU suppliers then said they could make a PC better and greener than any CPU supplier could and introduced the hybrid concept.

In other words—it’s hit the fan brothers and sisters.

The explosion

Whereas we had two basic architectures in the PC, we have now gone to eleven architectural possibilities.

1. AMD with internal memory manager and serial links.
2. Intel with FSB.
3. AMD with heterogeneous computing.
4. Nvidia with heterogeneous computing.
5. AMD with internal GPU.
6. Intel with internal memory mgr and serial links (rev 1).
7. AMD with hybrid.
8. Nvidia with hybrid.
9. Intel with internal memory mgr and serial links (rev 2).
10. Intel with internal GPU.
11. Intel with coprocessor.

And there are even more possibilities when we add dual AIBs, and AIBs with dual GPUs, and hybrid combinations of them.

In the past century, we saw the explosion of suppliers, a half dozen CPU suppliers, 70 3D graphics chip suppliers. Then we saw the consolidation to two and two.

In the last century we saw the seemingly unending escalation of clock frequency. Then we saw the sudden asymptote and the subsequent horizontal expansion of computing cores.

An d now those few suppliers, and all those cores, are being brought to bear on innovative and novel new architectures that are going to cause the OEMs and consumers’ heads to spin if not
explode.

Adding to it, we have a new war of words about where the best bang for the buck is—in the CPU or the GPU? And naturally both camps have their proofs for the answer.

Fantastic market opportunities

The differentiation possibilities and potentials for the chip supplier, the ODMs, and OEMs are enormous. We are going to see such an explosion of SKUs that BestBuy and Dixons are going to have to build bigger stores. Web pages will expand exponentially.

Reporters, analysts, and websites will have a field day explaining it all to the unwashed masses, investors and each other. Ad revenue will soar. Inventories will soar. The economy will jump five points as the PC industry offers a system for every possible situation, user, and usage case.

The messaging for all these new configurations is going to be most challenging to the OEMs. They’re the ones who have to translate the value proposition of each one into simple bite-sized four word sentences so the consumer doesn’t get scared. Yet another great market explosion—ad agencies and PR firms.

Man-o-man and mano a mano, 2008 is going to be a great year.

Resistance is futile—you should believe

That might beg the question, “what exactly is CG,” but I’m going to sidestep that discussion because it might distract from my main point, which is that the most important thing in CG is photons. Too easy? OK, the most important thing in CG is sending a hellofa lot of colored photons, big fat photons called pixels to your central nervous system via the eyes.

OK, still too easy. I guess I am going to have to define CG. Ready? CG is the suspension of disbelief about the origin of what you see, and if done correctly, what you feel (physically and emotionally); well done CG is like a Turing test.

So with that definition, what’s the most important thing? Tricking the eyes and the computer system behind them all the way to your toes. In fact, that could be the ultimate test of suspension of disbelief, when you look at a CG image and curl up your toes in either fear or titillation—we could call it the toe-curl test and rate it on a scale of one to 10 with 10 being that it takes 10 minutes or so for you to manage to uncurl your toes and get the circulation back.

I’ve been wandering around the world lately, giving talks about CG and I have three images I like to use to make the point.

Of those three images, and subject to how you’re viewing them (on a screen or the printed page), which ones do you think are real—i.e., a photograph and which are CG? Take your time, I’ll wait. Da da dum de dum. OK, ready? The one on the left is a raytraced image, the one in the middle is a raster image, and the one on the right is a photograph of a painting. I’d guess everyone got number three—the Hopper painting. Most people I’ve given this test to don’t get the middle one, and depending on how I lead the audience most don’t get the first one either. So that’s pretty good suspension of belief, no?

But here’s the deal. Some of you identified all three correctly. Now, maybe that’s because you have trained eyes like me, or you’re just smart. Regardless, you weren’t tricked, and that’s what CG is all about—tricking you. And if you’re sitting in the cockpit of a simulator and you don’t get tricked, when you’re in the real deal you get dead.

So what’s the most important thing in CG? Tricking the viewer. Period. And how do you do that? With lots of pixels that relate to each other and the mind’s eye. Can you be tricked on a 15-inch 1024 x 768 screen, or a 17-inch 1280 x 1204, or even a 20-inch 1680 x 1050 screen? No. But you can come close on a 30-inch 2560 x 1600, and two of them jammed tightly together can get you even closer. Now if you had two 30-inch screens, you’d need some hefty pixel pushers because you’re talking about 8.2mpixels at least 30 times a second and all the pixel polishing that goes on behind those displays to get you to curl your toes.

And that is the heart of the issue and what CG is all about—compensa­ting for limited bandwidth, compute cycles, memory, and database delivery to create a sequence of unique (no MPEG here folks) images that will, well you know, the toe thing. CG is trickery to compensate for lack of resources—and sometimes we actually make it happen. All of which leads us to Peddie’s first law—in CG too much is not enough, and Peddie’s second law—the more you can see the more you can do (as in suspend disbelief). And that, boys and girls, is what gets us out of bed in the morning—working the tricks.

Final word

We’re not supposed to resist suspension of disbelief. We’re supposed to embrace it. So, in the examples you’ve just seen, we’re really supposed to want to believe.