All Processors Feeds

AMD vs Intel showdown

Alex Cox — 1208448567

We’re not the first to be criticised by certain AMD supporters. They argue that we’re biased towards Intel because of the lower review scores we usually give AMD-based equipment. And yes, the review scores are lower, but there should be no question of any bias – we desperately want AMD to start succeeding.

Every processor launch, platform arrival or big announcement sees our ears perk up like a dog that’s heard its master rattling its bowl, but we’re perpetually left with our tail between our legs when we actually get our hands on the products to put them through our rigorous tests.

One of AMD’s press relations agencies contacted us to offer a straight comparison – one machine with Intel, one with AMD – suggesting enthusiastically that AMD’s mobile platform had been misconstrued, and it was synthetic benchmarks (those that don’t require user input, like 3DMark) that had shaped our opinion on AMD.

Run real world benchmarks on the machines, we were told, and we’d soon see the difference. Well, we privately guffawed at the thought of AMD’s Turion X2 architecture besting the venerable Centrino platform, and then gratefully accepted their offer.

After all, they wouldn’t have offered if there hadn’t been something in it, right?

Test 1: Boot

Real world test number one, then. One of the most annoying things about owning a PC is the time it takes to get going; what could be more real than that as an initial test? Nothing.

And AMD, sadly, has dropped its chips in a puddle on this one. From a complete shut down to the appearance of the flashing cursor on the Vista Basic log-in screen, the Centrino machine took 38 seconds, while the Turion bumbled away for an additional 19 seconds, coming in at 0:58.

It did, however, manage to pull some seconds back when we then tried a cold restart, taking 1:16 (against Intel’s 1:08) to make it back to the password prompt. These being laptops, you’re just as likely to be returning them from sleep mode as you are to be booting them up from cold.

The Turion managed a fully respectable 24 seconds when entering sleep mode from the Vista desktop, and an eerily similar 24 seconds to wake itself back up again.

We’d like to confirm that we think this is a perfectly fine lick of speed, just in case any sensitive AMD fans are reading. Thirty seconds would be fine. But when the Centrino manages to enter sleep mode in 10 seconds and wake up in 19, we have to give this to the Intel platform.

Winner: Centrino

Test 2: Battery

Continuing the real world theme, here’s some advice for aspiring laptop owners: don’t buy either of these machines if your intention is to use them on the move; not only are they monstrously heavy, they appear to have one rechargeable AA battery’s worth of portable power.

And establishing a new theme of using synthetic benchmarks while cackling at the thought that we wouldn’t, we enlisted the help of a wonderful benchmark called Battery Eater, which taxed the CPU, hard drive and graphics unit until each laptop gave up and shut down.

AMD’s machine lasted a paltry one hour 11 minutes, while Intel clung on to do one hour 34, but these extra 23 minutes don’t represent a real victory by any means – both scores are weak. The clumsily lit 17inch panel on these Inspirons was unlikely to have helped.

Winner: Centrino

Test 3: Graphics

Ah, here’s the ace in AMD’s hole – graphics. Collaboration with seemingly revitalised sister company ATI has seen AMD move towards a united platform model, where it creates processors, chipsets and graphics units that are actively designed to work well together.

Regardless of whether the plan works out in the end, it also means we’re seeing some excellent graphics technology being shoehorned into the chipset; the integrated 1270 inside the Turion machine should see AMD sailing through against Intel’s bog standard X3100 chipset. And it does! But not by much.

Cinebench (yes it’s synthetic, but we like it) rated the AMD’s OpenGL performance at 711 points, which Intel missed by a small margin. Visibly, though, the AMD results were much better; viewing the test with two identical screens shows that there’s a clear amount of extra processing power going on inside that ATI chipset.

Winner: Turion

Test 4: Gaming

With all the extra graphical goodness on AMD’s plate well and truly established, it was time to move back to the real world.

Or the bullet-filled world of paranormal shenanigans that is 2005’s FEAR, to be more precise; not the freshest of gaming morsels, but anyone expecting us to attempt a Crysis benchmark on these sub-£500 laptops has another think coming.

At the lower ends of the game’s performance settings, the ATI chipset retains its lead from the last round; all decadent and well-lit, with an average framerate in the 80s, it pummels Intel’s performance by a good 30fps.

Sadly, when pushing the graphics past the archaic 640 x 480 resolution, things start to go pear shaped for both machines; both manage scores around the 16fps mark when running in medium graphics mode and half that when running at high detail.

Despite the drop in framerate, though, the ATi card runs roughshod over this round. All this performance in the face of Intel’s undeniably hefty processor is a definite plus for AMD here.

Winner: Turion

Test 5: Price

Well, wouldn’t you know it? This heated contest has reached a final, deciding stage.

And it’s our unutterable pleasure to declare AMD the winner, since it undercuts Intel on price by nearly £50. But this is a slim victory; the Intel is the more responsive machine, the nicer machine to use day to day.

If you’re buying a lowcost laptop for work purposes, or just to browse the Internet, we’d recommend the slick mechanics of the Intel machine every time. But if you’re foolhardy enough to want to play three-year-old games on a bottom-of-the-barrel PC, you’re better off looking AMD’s way.

Problem is, you do have another option. Save up for just a little bit longer, put in an extra couple of hundred pounds, and grab a laptop with a better processor and meatier graphics on board, one which might even let you near some reasonably fresh games.

Neither of these machines is a gaming laptop; so the fact one of them does a reasonable job is neither here nor there.

Overall winner: Turion

Reviewed: AMD Phenom X4 9850 Black Edition

James Rivington — 1208274519

Ever since AMD became a major player in the processor market, things have been a lot more interesting. With both Intel and AMD pushing each other into building bigger, better and faster processors, the biggest winners were us, the consumers.

However, things have gone a bit pear-shaped for AMD in the second half of this decade. Despite acquiring graphics giant ATI, it's had its ass handed to it by Intel at almost every turn. Intel’s processors are quicker, they’re more power efficient and they’re more desirable.

And as a result, Intel is currently dominating the PC processor market with its Core 2 Duo and Quad chips.

However, that doesn’t mean AMD has given up; it’s hoping to claw back some market share with its new Phenom X4 9850 quad core processor. Can it succeed? And does the new processor get anywhere near Intel’s Quad core offerings?

Find out in our full review

Powerful SpursEngine ready for PC builders now

J Mark Lytle — 1207644535

You've heard all about the wonders of the mighty Cell processor that powers the PS3 and high-end industrial computers, so maybe now's the time to get in on the act yourself with Toshiba's Cell-based chipset about to go on sale.

The Tokyo company has announced that it is ready to ship samples of the SE1000 SpursEngine to manufacturers for the bargain price of ¥10,000 (£49) each. Mind you, there's a minimum order of several hundred units to qualify for the development support on offer, so a spot of DIY PC building might prove costly.

Upscaling video

Whatever the big PC builders do with the chip is sure to be impressive - demos we saw in Japan last year had a prototype Toshiba laptop upscaling standard video to HD TV and handling complex real-time graphics tasks.

The earlier-than-expected arrival of the SpursEngine promises consumer products powered by its four Cell-derived cores within a year or so. As for Toshiba itself, we're still waiting to hear if it has plans to bring a SpursEngine PC to market.

Intel's UMPC is still a crap idea

Jeremy Laird — 1207269208

The smorgasbord of technology, slightly crazed enthusiasm and side-stepped questions that is the Intel Developer Forum is over. Here's our take on the key highlights of IDF: Shanghai and what they mean for computing's future.

It's a long way to go...

At this IDF, Intel fleshed out speeds and feeds for the first ultra mobile variant of its Atom processor. But that was about it regarding significant announcements relating to Intel's core computer chip activities.

For that, thank a conference call a few weeks prior to IDF Shanghai that dished out details on both the upcoming Nehalem CPU architecture and the Larrabee graphics chip. Not enough news, perhaps, in return for the typically monumental IDF carbon footprint.

The UMPC is still a crap idea

Intel may have given the UMPC concept a new name - the Mobile Internet Device, or MID for short. But it's still the same lame idea. Indeed, when pressed to discuss the pocket-busting proportions of the MIDs shown at IDF, the typical reaction of Intel suits was to get a little shirty. We've a feeling Intel knows that MIDs are not the answer.

Splitting the Atom

In the long run, we've little doubt that future revisions of Intel's Atom Centrino mobile platform and CPU promise revolutionary performance and software compatibility improvements for mobile digital devices.

But if the oversized MIDs on show are anything to go by, the first iteration is simply too bulky and power hungry to change the world. We therefore suspect Atom will actually make more impact on traditional form factors. Expect to see some astonishingly cheap Atom-based desktops and notebook systems before the year is out.

Larrabee remains a mystery

Intel may have come clean with its intentions for the mystical Larrabee processor. We now know it is unambiguously targeted at Nvidia and AMD's graphics boards. In fact, Intel reckons it will kill those cards stone dead.

And there's little doubt Intel is hoping Larrabee will drag the entire graphics industry over to ray-tracing tech. But we remain clueless as to Larrabee's true potential in traditional raster-based 3D rendering scenarios. And that is what will make or break it, in its initial iteration at least.

Critical mass for solid state drives

Intel is nothing if not the absolute bomb when it comes to economies of scale in silicon chip production. So, confirmation that it is planning the imminent roll out of a full family of solid state drives (SSDs) looks awfully promising in terms of driving prices down.

In partnership with STM Microelectronics, Intel will shortly unleash SSDs spanning a range of capacities from 32 to 160GB. With any luck, the spinning hard disk will be dead for all but seriously high capacity bulk storage by mid 2009.

Political posturing

Is Intel shameless enough to go on stage with a straight face and talk about an energy saving partnership with China... all the while with the Chinese reportedly commissioning a new coal fired power station every week? You betcha!

Intel was also happy to give stage time to the jocular but somehow slightly sinister looking Chinese Minister for Railways. We're not sure about his presentation skills, but he certainly looked the kind of guy who can make the trains run on time. So, it was no doubt a great week for Sino-Intel relations. Just don't mention capital punishment or widespread human rights abuse, umm-kay?

Intel puts TechRadar on a par with The Economist...

...And The Wall Street Journal, The New York Times and BusinessWeek. At least that was the august company we shared (in our outgoing Tech.co.uk guise) in the form of a foil festooned with quotes from leading publications during the opening keynote on day one. No other technology publications made the grade.

Shopping in Shanghai

...And finally, don't go shopping in China unless you are prepared to haggle for Britain – don't be surprised to see the Chinese shoehorn in haggling as a demonstration sport at Beijing this summer. And if you do manage to beat your opponent down to your target price, for goodness sake don't then change your mind and flatly refuse any purchase at all. You'll be lucky to get out of his market stall alive.

If that makes shopping in Shanghai sound like a chore, try this for size. You can still have a suit made to measure and delivered to your hotel room within 24 hours. And all for around £70. Long live the Chinese experiment in capital-socialist fusion!

Intel's parallel processing vision

Ian Robson — 1207147034

With multi-core processors now established, what are Intel’s plans are for pushing the full potential of parallelism in computing? At the IDF in Shanghai, we caught up with Intel’s vice president and director of research, Andrew Chien, in an effort to find out.

Prior to joining Intel, Andrew Chien spent 15 years in the academic community working on various aspects of hardware and software for parallelism. Chien now leads many of Intel’s external efforts for research.

TechRadar: Will true parallel computing require a rethink, overhaul or a complete replacement of the current x86 CPU architecture that today's PCs are based on?

Andrew Chien: It’s safe to say that x86 is really deeply rooted in a whole range of software compatibility and toolchain aspects, which is the glue that binds the industry. When you look at parallelism, there are some issues in how you get things together and a whole new dimension of challenges.

We fully expect, because we’re Intel, that a full range of parallel systems will be built based on x86 or extensions of x86, and they’ll work just fine.

TechRadar: Have inherent difficulties materialised in the 30 years that x86 has been around?

Andrew Chien: Oh totally – but every instruction set architecture evolves from the time it’s introduced, and Intel has continuously been adding improvements to the architecture. After all, it started as an 8-bit architecture, right?

While most of the academic world has focussed on this RISC [reduced instruction set computer] versus CISC [complex instruction set computer] competition. In some parts of the academic community there is a deeply held belief that RISC won.

But the interesting thing is moving into [an age] where you’re frequency limited on the processor due to power. We’re seeing that now and Intel has changed its strategy. There’s now a lot of innovation around being more efficient from a power point of view with the instruction set.

So a lot of the extensions and the enhancements that you’re seeing coming out of Intel right now are really about that ability to describe a whole bunch of computation, with really more complex instructions. This allows you to achieve higher power efficiency.

TechRadar: Is this efficiency of power and the limitations on clock speeds also pushing forward the multi-core aspect?

Andrew Chien: Absolutely.

TechRadar: But beyond multi-core, will we also see a lot of operations pushed off the CPU and into the GPU or other discrete media processing hardware?

Andrew Chien: I think it certainly pushes modifications to the architecture extensions and so on that could increase power efficiency or single thread efficiency in a single stream.

The power offload is also a fundamental driver behind parallelism using multiple cores because it’s been known for at least 30 years that if you go parallel (and you don’t scale the clock as fast), you can get more operations per second, per Watt. It’s always just been so hard to go parallel as people haven’t wanted to go down that path.

This desire to offload stuff to special purpose engines was first seen with special engines for cryptography. There are also some media engines in small mobile devices, like a voice recorder. They have custom ASICs or custom designs bred to do the media encoding and decoding. Some of [this approach] is down to cost, but also some of that is because a hardwired design is more power-efficient compared to a general purpose core.

I wouldn’t really hold up GPUs as being any more power efficient. But I think there is a spectrum of general purpose designs, all the way down to hardwired implementations for power efficiency, which Intel is very conscious of. We look at integrating things into our SOCs [System On a Chip] and other kinds of products like that to address those markets.

TechRadar: The hardware developed for multi-core operations has far outstripped the rate of software development. What is Intel doing to get the software companies to code applications and operating systems that step up to the potential of the hardware?

Andrew Chien: In the long run, people need to write code that is fully scaleable, and frankly we need research breakthroughs to put the whole industry on that basis.

We’re calling upon governments as well as other players in the industry to invest in the five to 10 year future of scaling on parallelism. Because we need to move away from the traditional legacy of parallelism, which was just to get linear speed up, and to get maximum efficiency out of every CPU and every clock.

For parallelism to be successful we ultimately need to move to a world in which people write code that is inherently parallel; that won’t necessarily just get one more element of performance for every core that is added.

Intel is doing everything it can to create the urgency for this. I think one of the interesting challenges for parallelism and all application software is just where does it come from? Every time we’ve had one of these major changes in [processing] capabilities, often the largest consumers of the [processing] cycles turn out to be new applications.

In this space we’ve been out making people aware of a new class of workload called RMS – Recognition, Mining and Synthesis. It’s all about data streams and analysing large quantities of noisy data. It’s about finding insights from within [this data]; it’s about synthesising the whole graphic and the 3D visual experience.

These applications have staggering amounts of parallelism, so there’s no question that if those kinds of capabilities become increasingly part of other applications, then that alone could saturate many of these parallel processors.

This interview was conducted by PC Plus magazine Editor, Ian Robson

Intel showboats in Shanghai: IDF approaches

Jeremy Laird — 1206958375

The spring installment of the twice-yearly Intel Developer Forum kicks off on Tuesday in Shanghai, China. Apart from jet-lagged journos snoozing through the keynotes, just what can we expect from everyone's favourite silicon-spewing colossus?

More on Nehalem

This is Intel's bread and butter, its bottom line and its core competence rolled into one. PC processors, in other words, still define the company and Nehalem is the crucial next installment of its so-called 'tick-tock' CPU roadmap. Undoubtedly, it's the most significant architectural overhaul of Intel's X86 family since the original Conroe-vintage Core 2 processor blew us all away at the Fall IDF (I in 2006.

Intel has already lifted the lid on most of Nehalem's deep, dark secrets. We know that it will launch in native quad core format, pack an integrated memory controller and system I/O and even offer on-package graphics.

Oh, and Intel has also revealed Nehalem will resurrect Hyper-Threading technology – each core will be capable of crunching two software threads simultaneously. In short, Nehalem is shaping up to be a multi-threading monster.

The only thing we don't know is exactly how much extra performance it will bring. With any luck, that's something we'll get a pretty good feeling for at IDF. Intel might just allow us to test-drive the new chip in controlled conditions, just as it did with Conroe back in '06.

Until, then we'll go out on a limb and predict that Nehalem will bring only a relatively modest single-threaded performance boost – at most 15 per cent and probably more like 10 per cent, compared with the latest 45nm Penryn chips. Where the new architecture will really shine is in multi-tasking and multi-threading scenarios, particularly in bandwidth-intensive applications. Expect the fastest Nehalems to spank the best Core 2 Quads by around 30 per cent when threading is on the menu.

Menlow, Moorestown and masses of MIDs

X86 chips packed into some seriously small form factors should be a major theme in Shanghai this week. The Menlow ultra-mobile platform, otherwise know as Centrino Atom, is just about to hit critical mass in terms of viable production volumes. Expect Intel to rope in plenty of partners to help show off a slew of funky new mobile internet devices (MIDs) at IDF.

We'd also put good money on Intel indulging in a little trash-talking during its mobile keynote this time round. The target of Intel's wrath will be non-X86 mobile processors. Intel will be bigging up both the raw computer power and unbeatable software support, especially web-based applications, of its tried and tested X86 tech.

That said, Centrino Atom doesn't get really sexy until the second revision, known as Moorestown, arrives next year. Intel reckons it will deliver a 10x improvement in power efficiency. Moorestown will very probably be the first X86 platform to find its way into smartphones and other genuinely pocket-sized digital devices.

Larrabee

Intel keeps trickling out more and more details regarding the Larrabee multi-media co-processor, but it's still an infuriatingly mysterious entity. Is it essentially a graphics chip that can turn its hand to other tasks? Or is it a more general purpose floating-point beast that just happens to be a dab hand at pumping out pixels?

Hopefully Shanghai IDF will leave us with a better understanding of what the chip is capable of and exactly how Intel plans to position it.

Programmable silicon radios

Not a sexy topic on the face of it, but programmable silicon radios will have a huge impact on how you use information technology on the move. Intel is currently working towards a single-chip solution capable of supporting Wi-Fi, WiMAX, 3G, DVB-H mobile TV and much, much more.

Imagine all of those radios packed into a device as slender as the smallest GSM handset and you'll have some idea of what the future holds. IDF should bring us more details of the first step on the path to this fully unified ideal, in the form of a single chipset that combines Wi-Fi and WiMAX radios.

Even tinier trannies

No, not very small cross-dressers, but the inevitable march of Moore's law. Intel has been hitting its targets with unerring accuracy of late, so its upcoming 32nm silicon production node should be on track for full deployment next year.

It's a little early to expect any hands-on fun with 32nm kit at this IDF. But there's just a whiff of a chance that senior Intel suit Paul Otellini might briefly demo the 32nm Sandy Bridge die shrink of Nehalem on stage. OK, that's more likely to take place at IDF San Francisco this autumn. But it wouldn't be IDF without Intel showboating about its latest breakthrough in silicon wafer production.

Bendy chips to change tech forever?

Dan Grabham — 1206713617

The microprocessor might be an amazing thing, but durable it most certainly isn't. So what if a next-generation processor could be bent or stretched into unusual shapes? It'd certainly change things a little, wouldn't it?

Scientists at the University of Illinois have developed new materials to create flexible component boards that could be placed in places that circuits have traditionally not been able to go - the body and in clothing, for example. Of course, medical technology could also be used

"The notion that silicon cannot be used in such applications because it is intrinsically brittle and rigid has been tossed out the window," said John Rogers, professor of Material Science and Engineering at the University.

Rubberised chips

The findings, published in the US journal Science, revealed how the researchers have been able to create circuits with transistors and other elements from so-called nanoribbons of silicon crystals combined with plastic. The circuits are then backed with silicon rubber to make them malleable.

If stacked on top of one another, 10 of the 'boards' would be as thick as a human hair, say the researchers.

"We're opening an [opportunity] for electronics and optoelectronics that goes well beyond what planar configurations on semiconductor wafers can offer," added Rogers.

The material used may be malleable, but the scientists say the deal breaker is that there is no loss in electrical conductivity.

The research adds to the previous work carried out in 2005 by Professor Rogers and his team - they developed stretchable circuitry, but the pull could only be mono-directional.

AMD's new triple-core CPU and faster, cheap quads

Jeremy Laird — 1206611807

The wraps are officially off the new triple-core revision of AMD's Phenom CPU along with quicker quad-core models. Surprisingly, AMD said its new triple-core Phenom X3 remains saddled with the infamous bug that has affected the launch of the Phenom line.

Still, for what it's worth the tri-core layout of the Phenom X3 range is a first for PC processors. It's a core count currently not offered by Intel. As for the much needed revision of the quad-core desktop Phenom CPU, it not only brings faster clocks but also new circuitry tuned to tame the so-called 'TLB' bug.

The new chips are clearly a step in the right direction. But are they enough to turn around AMD's fortunes? One thing is for sure AMD is sure pricing these models very competitively - more on that later.

AMD's triple threat

First up is the Phenom X3 triple-core processor. AMD's Leslie Sobon confirmed the new processor is essentially identical to a quad-core Phenom chip, but with one core disabled. However, she flatly denied the suggestion that quad-core chips with one dud core were being used to knock up Phenom X3s.

If true, it's an odd state of affairs. Does AMD really intend to hobble perfectly good quad-core chips and flog them as triple-core? Wouldn't it make sense to increase its margins by using 75 per cent functional quad-core dies that will otherwise be consigned to landfill?

We may never know. Anyway, at launch AMD is announcing two Phenom X3 models. The X3 8400 clocks in at 2.1GHz and sports 512K cache per core and 1.5MB of shared cache memory. The 8600 cranks things up to 2.3GHz with the same cache counts. Both are rated as 95 watt chips and are compatible with most AM2 and all AM2+ motherboards.

Quad-core Phenom revisited

Spearheading the phalanx of new Phenom X4 (AMD says it has resurrected the 'X4' suffix due to customer demand) quad-core models is the 9850 Black Edition. As with previous 'Black Edition' chips, the 9850 has an unlocked multiplier for painless overclocking.

Speeds and feeds for this new range topper are 2.5GHz with 512K cache per core and 2MB of shared cache. Welcome though the boosted frequency is, it does come at a price. The 9850 BE is rated at 125 watts, a fair few notches higher than the new 9750 2.4GHz model, which is a 95 watt chip.

As indeed are the 9650 and 9550 models which round out the revised Phenom family. They're 2.3GHz and 2.2GHz quad-core CPUs respectively, by the way.

TLB still bugs AMD

Also worth scrutinising are the final two digits of the new processors' model numbers. The new '50' series Phenom X4s denote the latest silicon revision which, according to AMD, finally banishes the nasty little TLB bug present in all previous Phenom processors. The bug causes instability in certain situations and the firmware workaround AMD created to address it comes at a significant cost in terms of performance.

It's therefore a bit of a shocker to note the first triple-core Phenoms are buggy '00' chips rather than being based on bug free '05' silicon. AMD says '50' series Phenom X3s will come later.

What price success?

But what about pricing? This is where things get really interesting, not to mention worrying for AMD. The 2.5GHz Phenom X4 9850 BE is yours for just £115. The 2.4 and 2.2GHz models weigh in at £105 and £95 respectively.

As for the Phenom X3s, AMD is currently declining to divulge precise positioning. All we know is that they will be even cheaper than the X4 range.

In the short run, that's just gravy for punters looking for a good deal on a multi-core PC. But it must be absolutely decimating AMD's bottom line. Back in 2005, remember, AMD's top dual-core Athlon chip was changing hands for around £600. Even the relatively feeble Athlon 64 X2 3800+ commanded over £250. Today, AMD can't even manage half of that figure for its flagship quad-core processor.

Doubts remain

Question is, therefore, exactly how long can it afford to be restricted exclusively to the budget end of the CPU market? It would be a tough job for a well funded AMD to close the gap to Intel. But how is it going to fund the intensive R&D required to take the fight to Intel with the current firesale on all AMD CPUs?

We dare not speculate. In the meantime, AMD says the revised Phenom X4 family will be available “just after” today's official launch. Phenom X3s will be flowing in mid to late April.

AMD to downsize

1206009511

AMD is rumoured to be looking to lay-off 5 per cent of its workforce due to the economic downturn and the delay of its Phenom 9700 and 9900 processors.

The chip manufacturer, engaged in a long running battle with Intel for supremacy, announced in January that it was pushing back its eagerly awaited Phenoms to the second quarter of the year.

Combined with the credit crunch, this delay appears to have taken its toll on AMD, which is now strongly rumoured by The Enquirer to be looking to downsize.

Five per cent of the workforce equates to around 800 jobs, with the cuts expected to be across the entire company rather than aimed at a specific group.

Intel's Nehalem is a multi-threading monster

Jeremy Laird — 1205852930

Intel claims its upcoming Nehalem CPU architecture will deliver a major boost in multi-threading performance.

Speaking yesterday to hacks from across the globe, senior Intel suit Stephen Smith revealed a raft of new Nehalem architectural details. However, he conspicuously declined to make specific performance predictions.

Nehalem is the next step in Intel's much-touted "tick-tock" strategy that sees a new production process and major CPU architectural revision introduced in alternate years.

Late last year saw the launch of the first 45nm Penryn processors, essentially a die shrink of the 65nm Conroe family of 2006. Intel's schedule therefore demands that the Nehalem architecture rolls out on Intel's 45nm node before the end of '08.

More parallel processing

First up, Intel has boosted Nehalem's simultaneous micro-ops handling from 96 in Penryn to 128 per cycle. If that sounds like digital double-speak, the idea is an increase in instruction-level parallelism. That in turn means more work done per cycle and more number crunching grunt.

The wider micro-ops handling is also said to to help keep both instruction pipelines in each Nehalem core operating near peak efficiency. Lest you have forgotten, Nehalem sees the resurrection of HyperThreading. That's a feature last seen on Intel's defunct Pentium 4 processors. It enables a single core to crunch two software threads at the same time.

Among other detail changes, Smith also said the Nehalem branch predictor had been improved, an enhancement that reduces the likelihood of processing cycles being wasted.

A question of cache

The other major disclosure was a break down of Nehalem's cache structure. All Nehalem chips will pack 64KB of L1 cache and 256KB of L2 cache. Those are supplemented by a large pool of shared L3 cache, in the case of the quad-core model, 8MB. It's an approach that closely mirrors that of AMD's troubled Phenom processor.

Intel has already revealed that Nehalem will be available in configurations from two to eight cores. Factor in the wider micro-ops handling and HyperThreading support and you have a chip with monstrous multi-threading throughput.

Nehalem's integrated triple-channel memory controller and on-die northbridge have likewise been subject of previous disclosure. Both are expected to increase the bandwidth available to the cores and, again, give a big boost to multi-threaded oomph.

Intel goes thread happy

All terribly impressive. However, we can't help thinking Intel has gone a bit thread happy. Today's quad-core processors already feel like overkill in terms of multi-threading, for average users at least. It's hard to imagine many PC owners benefiting anytime soon from a chip that can handle eight or even 16 threads.

These new details will no doubt have CPU architecture aficionados chatting deep into the night. However, the truth is that we still don't know just how much extra performance the Nehalem family will bring. With any luck, that's something we'll learn more about at the Shanghai instalment of IDF in a few week's time.