From iconic machines like the IBM PC to the silliness of Sony’s £1,190 netbook, we’ve scoured the darkest corners and blown dust off some of the oldest, oddest and rarest kit we can find – starting with a true icon of the industry.

IBM PC

Introduced on August 12 1981, IBM’s Personal Computer was the first machine to popularise the now-ubiquitous term – and one of these antiques sits at the back of the PC Pro Lab.

Processing grunt was provided by the single-core, 4.77MHz Intel 8088, and floppy disks and cassettes are both supported. There’s a mighty 256KB of RAM, with 64KB of that soldered onto the motherboard. The IBM PC didn’t come cheap, either: a barebones model without any drives cost $1,565 and the top-end model came with bells, whistles and a monitor for $20,000.

The motherboard includes five eight-bit Industry Standard Architecture slots, with three of ours occupied: there’s a floppy disk drive controller card a SixPakPlus memory expansion board packed with 64KB chips, and a multidisplay adapter that’s actually two slabs of PCB stuck together. In the middle of the machine is an IBM 5 ¼in Diskette Drive.

Only one question remains, though, once we’ve blown the dust off this venerable old machine – can it run Crysis?

Apple Macintosh Plus

The Macintosh Plus might be a disturbing shade of yellow but that’s hardly surprising – it first saw the light of day in 1986. Released for £2,599, it was produced until October 1990 – the longest production run of any Macintosh – and was supported by Mac OS up to 1996.

It broke ground in other ways, too. As the first Macintosh to include a SCSI port it paved the way for external devices such as hard disks, tape drives, printers and CD-ROM drives, and this was also the first Macintosh to use SIMMs for its memory – with a massive 1MB of the stuff included as standard across four 256KB sticks.

Our particular model bears the familiar Cupertino, California label on its rear, but the sticker also reveals that this machine was “Assembled in Ireland” – a far cry from today, where most technology seems to be produced in Asia.

Apple Macintosh Colour Classic

Fast forward a few years – and look under a different test-bench – and you’ll find another piece of Apple history. It’s the first compact Macintosh computer to come with a colour display, and we wouldn’t have the iMac – the world’s finest all-in-one PC – without the Colour Classic paving the way.

Originally priced at $1,400 in February 1993, it ran on Mac OS 7.6.1 – the first version of the OS to drop the “System” from its name so the more distinctive moniker could be trademarked and the OS licensed to third-party Macintosh manufacturers.

This ancient all-in-one was more upgradeable than most of today’s models, too. The Processor Direct Slot was used with the Apple IIe Card, and ran software designed for the older Apple II. This backwards compatibility was supposed to entice the education market to upgrade from Apple II machines to fully-fledged Macintoshes, but other upgrades were also available, from CPU accelerators to Ethernet and video cards.

This versatility means the Colour Classic enjoys a cult following today: users have modded the machine with Power Mac parts so its screen runs at 640 x 480 rather than 560 x 384, and others have fitted motherboards from more powerful models.

Casio Cassiopeia E-115

Technical editor Darien Graham-Smith found the Casio Cassiopeia E-115 hiding at the back of his cupboard, but it first arrived back in October 2000 when PDAs, rather than smartphones, were big news.

So, what did you get for £422? There’s the sturdy exterior, which we described as “dull-grey silver” and “resting on its laurels”, alongside a cradle that “feels cheap and doesn’t engage with the Cassiopeia as solidly as we’d like” in its full review.

It wasn’t all bad news, with a 240 x 320 LCD screen that was better than its rivals, and a 131MHz StrongARM processor that was “fast enough to ensure instantaneous contact searches and speedy application switching”, according to us. It also had 16MB of ROM and 32MB of RAM memory – “about as much as you currently need”, at least back then.

Oh, and the software? Microsoft Windows CE 3.0 PocketPC Edition. Our model is old and, presumably, scarred by Darien’s cupboard, so it wouldn’t turn on – although that’s probably for the best, given that we concluded that the Casio simply couldn’t “match the standard” set by Compaq’s iPAQ.

Apple iMac G4

The G4 marked the first major redesign of the iMac, but the forlorn model found in the PC Pro Lab has clearly seen better days. It’s missing its monitor bezel, the distinctive round base is looking grubby, and it wouldn’t turn on – although that chrome, cantilevered arm is as smooth as it was when the G4 was eased from its box in 2002.

We described the G4 as “smooth and elegant design that puts other computer makes to shame”. Even now it stands out in a sea of modern all-in-ones that all look a little too familiar.

Our review also highlighted Apple’s concentration on “excellent design and ease of use”, but that has downsides – a specification we described as “Paleolithic”. It’s the first time we’ve seen computers compared to dinosaurs, but the SDRAM was slow and the GeForce 2 MX graphics chip was a generation behind the curve. It might look nice – as Apple devices are wont to do – but PCs ran our Photoshop 7 benchmark almost twice as quickly.

Dell Latitude

The oldest laptop we managed to find demonstrates the changing of technology. This Dell Latitude isn’t quite as backward as we first thought. It’s either a C540 or C640 – we’re not sure which, as it’s been hidden on a high shelf for far too long – and it’s a mix of old problems and forgotten boons.

It’s running a Pentium 4 chip with Windows XP, but the most striking thing about this machine is its design – or lack of it. Plain plastic is the order of the day, and the lid features the familiar Dell logo, along with the kind of build quality that we’d slate if this machine were reviewed today.

The base doesn’t cover itself in glory, either, with stickers, flaps, screws, feet and even some exposed fans. It’s also obvious where laptops have fallen backwards as companies rush to build slim, snazzy Ultrabooks: we rarely see keyboards with the kind of comfort, responsiveness and travel as this Latitude offers, and the 4:3 screen has a native resolution of 1,600 x 1,200 – a huge amount of desktop real estate compared to the 1,366 x 768 and 1,600 x 900 screens that now seem to be the norm.

Nvidia GeForce 7300 GT, 7600 GS and AMD Radeon HD 2600 XT

We’ve a big plastic tub full of graphics cards in the Labs and, while most of them are recent, a trio of PCBs lurking amid the anti-static bags and DVI to D-SUB adapters come from decidedly older stock. Two Nvidia cards, the GeForce 7300 GT and 7600 GS, are joined by AMD’s Radeon HD 2600 XT.

They were launched in 2006 and 2007, and they handily illustrate the impressive speed at which technology is pushed forward. The first 28nm GPU has just arrived but, back then, Nvidia and AMD were using 90nm and 65nm processes – and the 390 million transistors in the AMD card pales when compared to the 4.3 billion in AMD’s latest.

The bandwidth statistics are telling, too: the Radeon card churns through 35.2GB/sec in its 512MB incarnation, with the 7300 GT and 7600 GS offering 10.67GB/sec and 12.8GB/sec respectively. The latest high-end card, the Radeon HD 7970, chews through 264GB/sec – and even modest boards, such as Nvidia’s GeForce GTX 550 Ti, handle 98.5GB/sec.

Oh, and our benchmarks? The 7600 GS played Call of Duty 2 at 18fps when run at 1,280 x 1,024. Bless.

Sony VAIO P-series

Sony senior vice president Mike Abary famously said his company would never join the “race to the bottom” when netbooks hit the big time, and he wasn’t joking – Sony’s VAIO P-series cost £1,190 inc VAT for the top-end model.

That money paid for radical design, with a base occupied entirely by the keyboard,  that’s still so small and fiddly that you have to peck at the keys, prod at the trackpoint and squint at the 8in 1,600 x 900 screen. The Z-series Atom was decidedly Z-list, too, thanks to performance that couldn’t match £350 rivals.

ARM and Nvidia are major mobile players, so it pays to listen when the firms announce new technologies. The latest developments from both serve up an interesting similarity with regards to how these companies are tackling one of the biggest annoyances of the modern smartphone: inefficient batteries.

ARM’s recent announcement, big.LITTLE, pairs one of its high-end Cortex A15 MPCore chips alongside an entry-level Cortex A7, which consumes much less power. It’s designed to seamlessly takes over when a device is tackling low-intensity tasks, so the power-sucking A15 is reserved for intensive games and apps.

Nvidia’s Tegra 3 chipset, meanwhile, uses a near-identical approach: the four cores on the main chip can be turned on and off to tackle everything from basic web browsing to high-end games but, if the phone’s in standby mode or you’re running low-power apps, those four cores will shut down, with processing power provided by a “Companion Core”. It’s based on the same Cortex A9 used by the main Tegra 3 chip but, crucially, it runs at 500MHz instead of 1.4GHz.

They’re both designed with power saving in mind – in fact, low-power processors are one of the main pillars of ARM’s business – but the fact that these workarounds have to be considered at all merely serve to highlight just how little innovation occurs in the battery market.

It’s clear that lithium-ion power packs just aren’t keeping up with the demands of modern devices, with more powerful processors and bigger screens hampered by batteries that aren’t much bigger or more efficient than those included in phones from a couple of years ago. While components, screens and apps have become more sophisticated and impressive, batteries are stuck in a rut.

It doesn’t seem like the problem will be resolved any time soon, either. Nvidia and ARM might be tacking on chips to handle low-intensity tasks but, as soon as you ramp up four Tegra 3 cores, your phone’s longevity is likely going to plummet, especially if it’s used on devices such as the Samsung Galaxy Note, which has a whopping 5.3in screen.

One of the few drawbacks of modern smartphones is the daily charge: a ritual that sees people leaving phones attached to the mains at night or to their computers at work. Until a newer, more efficient type of battery arrives, this will continue to be the case – even if Nvidia and ARM are doing their best to give outdated lithium-ion hardware plenty of help.

I’ve become used to Nvidia and AMD rocking up with new graphics cards on a consistently quick schedule: since 2008 and beginning with the GeForce 9000-series, Nvidia has released six generations of GPUs, and AMD has replied with three of its own, starting with the Radeon HD 4000-series.

Both firms put the finishing touches on their current ranges in the spring, and the trail’s gone quiet since. Nvidia has confirmed that it won’t release any new cards before 2012, and AMD hasn’t mentioned any potential release dates for its new Radeon HD 7000-series — presumably taking extra time to further tweak and perfect the new chips.

I can understand why Nvidia and AMD aren’t in any hurry to unveil new silicon — after all, evidence suggests that few games, and fewer gamers, will make use of the levels of power available from new cards: 2007’s Crysis is still one of the most demanding games around and, outside of Battlefield 3, few forthcoming games look like they’ll tax even last year’s graphics cards — after all, most are now ported from consoles.

Take a look at the Steam Hardware Survey from July 2011. The most popular ten graphics cards are, with no exceptions, old: top of the pile is the Nvidia GeForce 9800 GT, and it’s followed up by venerable old warhorses like the GeForce 8800 GT, Radeon HD 4870 and even the GeForce 8600 GT. The most popular current-generation chip, the GeForce GTX 560, is 25th on the list.

To find something truly high-end, keep on travelling down the list — the GTX 570 is the 31st most popular card and is used by a whopping 0.78% of Steam’s users, and the AMD Radeon HD 6950 is favoured by just 0.68%.

That’s the biggest PC gaming platform in the world and proof that, while AMD and Nvidia have spent the fast few years trying to beat each other with benchmarks, most users haven’t taken any notice. Instead, they’re too busy playing games on cards that are older and perfectly capable.

Perhaps the big guns have realised that this pixelated arms race is virtually pointless. Almost all users either aren’t bothered or can’t afford the latest chips, and the lack of big-name PC exclusives means that even a modest card will run rings around the latest games. Enthusiasts might enjoy the yearly releases, but board partners constantly tweak, overclock and improve chips to eke more performance out of older cores — and, if they’d like to run games across several screens, there’s always CrossFireX and SLI.

It looks like we won’t see any major graphics card releases until 2012 but, after several years of frenzied, benchmark-driven battling, I’m happy to wait — and it looks like most users are, too.

Lining up the systems on our test benches, though, proved just how spectacular these PCs are: SSDs, water-cooling and touchscreen fan controllers on the inside; triple monitors, adjustable mice and £400 speakers on the outside.

As well as benchmarking, I’ve spent plenty of time in the Labs taking pictures of these stunning computers. To get the full low-down on which system takes the Ultimate PC crown, you’ll have to wait until Thursday when PC Pro issue 204 hits the shelves. Until then, take a look at these pictures, try not to drool, and pick out which you’d buy if you had several thousand pounds to spare.

Remember AMD Barcelona? Delayed and disappointing, the architecture behind AMD’s first Phenom chips finally turned up in November 2007, six months behind schedule, with performance that left us “a little underwhelmed” after months of anticipation – and that’s before it was compared against its Intel equivalents.

At the same time, Intel was preparing to release its Wolfdale-based Core 2 Duo processors, which appeared in January 2008 using the more efficient 45nm architecture – a key improvement over AMD’s 65nm chips. The result? Our review concluded that the new Core 2 Duo E8000-series “wipes the floor with the [older] E6000 series” and that Intel’s new processors were an “unqualified success”.

Fast forward three years, and the similarities are startling.

The chips are down

Intel has recently unleashed its stunning Sandy Bridge architecture, which combines improved efficiency with markedly improved performance over its predecessors. The launch might have been marred by the firm’s well-publicised SATA problems, but they’ve now been fixed.

AMD, in the meantime, is floundering: its latest processors are retooled versions of old hardware, and forthcoming Bulldozer chips look like they’ll have an uphill battle, too. AMD is in second place when it comes to implementing 32nm manufacturing, with Bulldozer-based desktop processors not due until later this year and laptop parts only slated to arrive in 2012.

And then there are Bulldozer’s headline features, many of which sound awfully familiar. The chips will be built around individual modules that contain two processing cores with Multi-Threading and Turbo Core technologies, so they’ll be able to address two tasks independently and overclock at will, too – so the firm’s top-end consumer chips will include four modules with eight cores.

Bulldozer might not be as efficient as Intel’s chips, either. AMD’s new chips will have a maximum TDP of 125W, with Intel’s most powerful Sandy Bridge chips coming in with a top TDP of only 95W. That suggests great power efficiency, less heat, and more overclocking headroom.

By the time AMD releases its first Bulldozer-powered processors – rumoured to be in the summer – Intel’s latest Core i7 Extreme chips will also be on the horizon with eight cores and, consequently, 16-thread support as standard, and using a more efficient manufacturing process. They’ll likely be more expensive than AMD’s priciest Bulldozers but, if recent form is to be believed, they’ll also be far quicker, further establishing Intel as the chip of choice for the enthusiast.

The picture isn’t much rosier in AMD’s previous stronghold at the budget end of the market. Fusion is taking aim at Atom but, even though we’ve found it to be a better performer than its rival, we’re not exactly enthused about its prospects. Our review described it as “a year too late” to make a real impact. The desktop chips still aren’t due for several months, and the sheen that surrounded netbooks back in 2007 has faded as CULV laptops, smartphones and tablets – all of which are powered by non-AMD technology – have gained in prominence.

Nvidia fights back

At least AMD has spent the past couple of years delivering market-leading graphics cards, with the firm often finding the perfect balance between price and performance. The tide’s starting to turn here, though, with Nvidia’s GeForce GTX 460 and its successor, the GTX 560 Ti, proving to be better prospects than AMD’s own HD 6950, HD 6970 and HD 6870.

The firms are now trading blows in the high-end space, too, although Nvidia’s likely got the edge here, with its GTX 570 offering the “best compromise between cost and power”. And, while AMD currently holds performance records thanks to the new HD 6990, Nvidia is rumoured to be releasing its own dual-GPU card in the next few weeks – so you’d be wise to hold off if you’re thinking about dropping such a vast sum of cash on a graphics card.

There’s turmoil at the top, too. January saw the departure of CEO Dirk Meyer, allegedly because of the firm’s failure to capitalise on the growth of mobile devices, and other key staff left not long after, with COO Bob Rivet and senior vice president of corporate strategy Marty Seyer both announcing their resignations at the start of February.

In the meantime, AMD has appointed a couple of new faces to its board of directors. Henry Chow and Nicholas M. Donofrio, both veterans of IBM, bring plenty of chip-designing experience to the table – but, crucially, Meyer hasn’t been replaced, with Thomas Seifert manning the fort until a permanent CEO is appointed. That can’t come soon enough – while Meyer has an engineering background and personally led the team that developed the Athlon processor, Seifert is a manager with less technical expertise.

It’s a bleak picture. AMD doesn’t lead the way in any area at the moment, and few signs point to this situation improving. Beating rivals such as Intel and Nvidia might be one step too far for the Californian company right now – by the looks of things, it’ll have a hard enough time trying to keep up.

Several other details can be surmised by looking at these pictures closely, though. It’s brimming with display outputs — one DVI-I port alongside four mini-DisplayPort outputs — and the mammoth cooler extends down the length of the huge PCB and is topped off by a fan in the middle of card.

Of course, the HD 6990 might need that amount of cooling, seeing as it’s rumoured to include two of the firm’s latest GPUs.

That’s all the information we’re allowed to reveal for now, though, so you’ll have to check back soon when the embargo lifts for our full review. In the meantime, we’ll see if we can find a case and power supply that will accommodate this gigantic monster.

To read our full review of the HD 6990, and find out if this £550 card is worth the cash, click here.

The leaked information concerns the Radeon HD 6750 and HD 6770 which, if the past two generations are to be believed, will sit in the middle of the upcoming range. There’s evidence to suggest that the new series is more evolution than revolution, with both cards still using the 40nm fabrication process that was introduced way back with the HD 4770 and the GDDR5 memory that’s been commonplace for the past year.

Nevertheless, the list of specifications hints at the increased power that AMD has been able to eke out of its new Northern Islands family, of which the Barts XT core is the first representative. The HD 6750 will allegedly have a 725MHz core accompanied by 1,120 stream processors, and its compute performance of 1.624TFlops sits between the HD 5770 and HD 5850 in the pecking order.

The HD 6770 is, potentially, more interesting. Its alleged compute performance of 2.304TFlops is higher than the 2.088TFlops performance of the HD 5850 and, with 1,280 stream processors on board, it’s sure to pack some serious muscle.

The HD 6770 even manages to beat the HD 5870 in a couple of areas, with a higher clock speed – 900MHz to 850MHz – and a higher pixel fillrate, with the HD 6770 managing 28.8GPixel/sec to the 27.2GPixel/sec of the older card. Its GDDR5 memory is even clocked 50MHz higher than the HD 5870.

So, on paper – and assuming this leak is accurate – the HD 6770 is likely to outpace the HD 5850, which is already faster than anything Nvidia can offer at that price, and could even serve up performance that approaches the HD 5870. And, if AMD’s prices will remain in the same ballpark, the HD 6770 will cost around £110 exc VAT; the HD 5850, meanwhile, cost around £170 exc VAT on release, with the HD 5870 north of £200 exc VAT.

Of course, this is pure speculation, but there’s no reason to doubt that AMD’s recent form will carry over into its new generation of cards, even if there’s little revolutionary about the architecture behind these parts. Nvidia, meanwhile, is still filling out the bottom end of its GTX 400-series of GeForce products with cards like the GTS 450, and looks to have little in the way of an answer, with rumoured price cuts the only feasible short-term option.

Nvidia has recently lost its majority share in the desktop graphics market to AMD, and it looks like this gap is set to widen. No wonder Nvidia’s concentrating on Tegra and mobile these days.

Any qualms as to whether Nvidia’s Tegra 2 is up to the job are soon dispelled – the Folio 100 eagerly zipped through the familiar Android menus, and the whole experience felt suitably slick and refined. The full-sized HDMI socket on the device’s edge gives a clear indication of the Tegra’s mighty graphics crunching power: this is one tablet that’ll relish the challenge of HD video playback. And, with Flash 10.1 on board, YouTube HD footage is only a few clicks away.

While most of the Folio 100s were tethered to a wall, secreted from the clumsy, prying hands of eager journalists, we found Toshiba’s conference demo unit sitting quietly in a dimly lit corner. But while we were pleasantly surprised by the fine figure of ViewSonic’s ViewPad tablets, Toshiba’s effort is a touch disappointing.

The 14mm thick chassis feels lightweight, and even relatively gentle twisting motions left the Folio’s plastic body creaking under the stress. It might weigh a very reasonable 760g, but if you’re expecting a Toshiba-branded iPad, this just isn’t it.

That 10.1in multitouch screen boasts the usual 1,024 x 600 pixel resolution, but image quality is also far from stellar. Tilting the Folio from side to side revealed narrow viewing angles, and the Android home screen looked noticeably less sharp than Viewsonic’s efforts.

At least the specification is more promising. The 16GB of internal storage is complemented by a full-sized SD card reader which supports cards up to 32GB. Wireless networking stretches to 802.11n speeds, while Bluetooth and 3G comes as standard. A 1.3 megapixel webcam rounds off the proceedings.

We’re not overly keen on the plasticky-feeling chassis, but with Toshiba suggesting the Folio 100 is going to retail for around 399 Euros, we’re willing to cut it some slack. The Folio 100 might yet be the Flash-enabled iPad killer some people have been waiting for.

After all, Nvidia has taken quite a beating from ATI, which has released two generations of GPUs in the time that it’s taken Nvidia to unveil its first new architecture since the G92 core, which ushered in a bevy of GeForce 8000-series chips. And then 9000-series cards. And then, after that, a raft of low-end GTX 200 and GTX 300-series parts.

It’s odd, then, that Nvidia has launched its new cards in such a low-key fashion. The official launch was at the Nvidia-sponsored PAX East event, which isn’t a trade show or celebrity-studded party but a gaming festival and LAN event held in Boston. While it’s a big event, it’s not on the level of CeBIT or CES, and there’s no sign of the concurrent global launches that numerous tech firms, including ATI, have favoured in recent years.

Instead, Nvidia’s presentation occupied the Main Theatre in Boston’s Hynes Convention Centre, but it hasn’t secured the top spot. That’s reserved for a keynote from Wil Wheaton – that’s Wesley Crusher from Star Trek: The Next Generation, sci-fi fans – and a concert of video game music. Other presentations across PAX East’s four stages include “Organising a Child’s Play Fundraiser”, “MMO Gamer Behaviour 101” and “Community Managers: More than Forum Monkeys”.

Other aspects of the launch seem strange. The timing, for instance: why has Nvidia lifted the embargo on such an important product at 11pm on a Friday night? There’s no doubt that the time difference does the European press no favours, but it’s still late afternoon in America, too.

If press officers want to bury bad news, they normally do it late on Friday, safe in the knowledge that many journalists will miss it and  fewer people will be reading websites over the weekend. It’s a very strange time to launch your new flagship product.

Is it possible, then, that Nvidia has buried the launch of its new card? Our review reveals that it’s not exactly the GPU we’ve been looking forward to – we found it only marginally quicker than the older, cooler and quieter ATI Radeon HD 5870 and around £140 more expensive.

The benchmark results suggest that Fermi isn’t as good as Nvidia hoped, and there’s certainly an air of desperation around the launch. Nvidia has constantly asked journalists to bear in mind the future potential of the card instead of current benchmark results, and there’s also the usual trick – of which ATI is equally as guilty – of manipulating performance graphs: rather than starting at zero, plenty begin half-way up the scale, therefore exaggerating any performance differences between Nvidia and ATI chips.

It’s also telling that, while Nvidia may have released the card on a Friday night, that product won’t likely be available for a couple of weeks: we’ve spoken to several of Nvidia’s board partners and they’ve confirmed that they won’t be shipping chips until 6 April. Until recently, some even confessed that they didn’t know when they’d be in a position to offer cards, because Nvidia hadn’t delivered reference boards or given any indication as to the launch schedule.

The LAN-party launch, lack of fanfare and inconvenient timing all point towards one conclusion: Nvidia knows that its new card is barely better than ATI’s cheaper, older offerings, and it’s in full-on damage limitation mode. Let’s hope that Nvidia’s mainstream chips – which will inevitably be released in the coming months – are better value than the two disappointing high-end parts that we’ve seen so far.

Oh and, in case you missed it at 11pm on Friday night, here’s our Nvidia Fermi review.

Actually, that might explain a lot. But I digress.

The point is that Nvidia, unlike Intel, is acutely tuned in to what people are saying about it — and can be quick to respond.

I well recall how, at last year’s Nvision event, one fledgling journalist received a stern dressing down from PR director Derek Perez mere hours after she’d posted an online article that cheekily – but accurately – reported his impression of CEO Jen-Hsun Huang’s keynote address. (For the record, it was “dull and boring.”)

So I wasn’t wholly surprised when this morning, at the conclusion of my meeting with CUDA general manager Sanford Russell, I was ushered into a luxurious suite on the twentieth floor of San Jose’s Fairmont Hotel for an impromptu chat with Tony Tamasi, Nvidia’s Senior VP for content and technology, on the subject of my last blog post.

A billion dollars on CUDA

It turned out that Tamasi wanted to respond to two points. The first was the doubt I had expressed over whether CUDA could ever be a real money-maker for Nvidia.

“Supercomputing,” he assured me, “is a billion-dollar market.”

This I could not deny; but given CUDA’s apparent focus on academia, it seemed a surprisingly ambitious figure for Nvidia to be bandying about.

So I asked: “Have you made a billion dollars from it this year?”

“Of course not,” Tamasi laughed. “But we believe in the potential. We’ve been investing heavily in that for years. And when the market arrives, we’re going to be at the head of it.”

He showed me a slide demonstrating how Nvidia expects its investment to pay off:

— and I could only agree that – if PowerPoint was to be believed – there did appear to be a lot of money out there for the taking.

“So you’re going to sell ten thousand GPUs to the Department of Defence?”

“At least!” he declared confidently.

There’s no telling how much this market will really turn out to be worth to Nvidia. But the company’s investments in research have indeed positioned it well to “trickle up” into industry and government; and with CUDA now programmable in C++, the company’s ambitions are sounding increasingly credible. I wouldn’t bet against the technology growing quickly in these areas – in the short term, at least.

Sizing up Larrabee

But what of the longer game? I was half joking when I suggested that Larrabee might displace CUDA, but Tamasi agreed that it was a possibility.

“Intel is… not very excited when they see a researcher talking about porting code from Intel CPUs to Nvidia GPUs and getting a hundred-fold speed-up,” he predicted.

“And those are the super-high-margin juicy CPUs for Intel.”

“So Intel is defending their computing front. And I agree with you that Larabee is at least partly an effort to try to keep applications from going to the GPU style of parallelism.”

Is Nvidia worried about the long-term challenge?

“Nobody knows how good Larrabee is,” Tamasi mused. “Probably Intel doesn’t know how good Larrabee is. So we take them incredibly seriously.”

“There are strengths and weaknesses to their style of architecture. And I just don’t know how that’s going to play out.”