ZX Spectrum: 30 years old, and still one of the cheapest computers ever made

Today is the 30th birthday of the ZX Spectrum, one of the most popular home computers ever made, and probably the single most important factor in the creation of the IT industry in the UK. The ZX Spectrum, made by Sinclair Research in Cambridge, England is usually considered the UK equivalent of the US-made Commodore 64.

Hardware-wise, the ZX Spectrum was completely unremarkable. There was an 8-bit Zilog Z80A CPU, a graphics chip capable of outputting 32 columns by 24 rows (256x192px) with 15 colors, and either 16 or 48KB of RAM. At just £125 ($200), however, the ZX Spectrum was incredibly cheap. The Commodore 64 cost $600. The BBC Micro, made by Sinclair’s arch rival Acorn Computers, cost £299. Despite costing just a fraction of its contemporaries, the ZX Spectrum had comparable functionality. All three computers had similar amounts of RAM and processing power, and all three had similar editions of the BASIC programming language.

How did Sinclair Research pull it off? Innovative design and aggressive engineering. From the very start, Sinclair Research knew that it wanted the ZX Spectrum to be as cheap as possible, and so almost every component was engineered from the ground up with penny pinching in mind. The main printed circuit board was kept as small and dense as possible, which resulted in a very lithe chassis (just 23x14x3cm, compared to the monstrous 40x21x7cm Commodore 64 and gargantuan 40x35x8cm BBC Micro). Instead of using a conventional keyboard with hundreds of moving parts, a rubber, chiclet “island” keyboard with just four or five parts was used. (In the eyes of original users, this resulted in the ZX Spectrum keyboard feeling like “dead flesh” — an early example of a tech meme.) The ZX Spectrum was wrapped in a plastic case and weighed just 550 grams (1.2lbs), compared to the metal, clunky 1.8kg (4lb) Commodore 64, and back-breaking 3.7kg (8.1lb) BBC Micro.

In short, the ZX Spectrum was simply better engineered than its contemporaries — much like iPhone, except Apple uses its engineering and supply line advantage to squeeze out higher profits, rather than slashing prices. Like the ZX Spectrum, it’s not like the iPhone uses fundamentally different silicon or materials — Apple is still limited by the state of the art — but through design, engineering, and supply line expertise, Apple simply manages to cram more tech into the same (or smaller) space — and with a cheaper bill of materials.

ZX Spectrum+, a later version that did away with the “dead flesh” keyboard

The ZX Spectrum would go on to sell five million units — not bad, when you consider there are only 30 million homes in the UK — and net Clive Sinclair, the owner of Sinclair Research, a knighthood for “services to British industry.” Curiously, Sinclair, a serial inventor, recently admitted that he doesn’t actually use computers — he prefers the telephone to email.

To this day, even after 30 years of being hammered at by Moore’s law and accounting for inflation, there are remarkably few home computers that have been sold at a lower price point than the ZX Spectrum (it would cost around $450 today). The Raspberry Pi, a British-made Linux-based PC that will be sold for around $25, is the obvious exception, and the spiritual successor of the ZX Spectrum.

Source

Intel announces new smartphone SoCs, aims for emerging markets

Even back when Intel first unveiled Medfield at CES, we knew the company was planning a number of additional announcements for Mobile World Congress this week. Things kicked off yesterday with UK provider Orange’s announcement that it’ll be rolling out Medfield phones this year. Orange isn’t a company US readers will be very familiar with, but it’s a major player in the UK and France with over two hundred million customers. The new phone will offer 16GB of storage, a 4-inch screen, and weighs in at 117g (~4 oz). The one significant downside is that it’ll debut with Gingerbread; those eager for ICS will have to wait for an over-the-air update promised for not long after launch.

Orange is just one of several partners announcing today, and new partners are only part of what Intel is unveiling. When we covered Intel’s new Z2460, we told you that the new chip would be clocked at 1.6GHz. Intel has since adjusted the Z2460′s maximum clock speed upwards by 25%; the chip will now clock up to 2GHz. Which OEMs will take advantage of this option is less clear; Intel cautions that “each customer implementation is different.” Whether the current crop of devices use the higher 2GHz threshold or not, Intel’s ability to ship a faster clock speed says good things about the chip’s yields and overall power consumption.

The Z2460 isn’t the only smartphone SoC Intel is launching in 2012. The second half of the year will see the debut of the single-threaded Atom Z2000 and the Intel XMM6265 modem. The Z2000 will be a low-power Atom variant clocked at up to 1GHz; the XMM6265 is an HSPA+-capable modem. The current Z2460 Atom SoC is paired with the XMM 6260; it’s unclear if this new chip is a simple tweak to that product or a new, low-power 28nm variant.

One of the hotly debated topics in the wake ofMedfield’s launch has been whether or not Intel can build a chip to fit into a low-power form factor. The reason that it can — and why a 1GHz Atom phone will draw minimal amounts of power — is highlighted in the graph to the right.

This graph shows the chip’s power consumption at various clock speeds. Low frequency mode’s (LFM) operating clock speed is well below 1GHz, while HFM is moderately above it. As a result, a 1GHz single-core Atom should draw very little power. Unlike its bigger brothers, the Z2000 won’t offer Hyper-Threading, but that won’t be a problem given the markets the chip is aimed at.

Intel’s first high-end chip in the smartphone market will be the 32nm Atom Z2580; it’s currently scheduled to debut in early 2013. The Z2580 will be a dual-core variant of the current Medfield with two cores clocked at 1.8GHz with Hyper-Threading enabled and a dual-core GPU based on the SGX544 at 533MHz. The Z2580 will also launch with a new, LTE-capable modem, the XMM 7160. Intel’s current high-end radio, the XMM 7060, is also built on a 40nm process, but in this case, we’re willing to bet that the XMM 7160 will be a 28nm variation of that part. Current 40nm LTE radios draw a tremendous amount of power; all smartphone vendors are talking up their transitions to 28nm designs.

The SGX544 is the fully DX9-capable version of the SGX543 (found in the iPhone 4S and iPad 2) and is twice as powerful as the SGX540 at the heart of Medfield. The Z2580′s GPU will also be clocked 33% higher than the Z2460′s, with two cores running at 533MHz as opposed to a single 400MHz core. As with Medfield, Intel is pushing current tech to higher clock speeds to counter other, potentially more advanced architectures — the generally quoted fillrates and GFlop figures for the SGX543/544 family assume the chip is running at 300MHz.

Of the two new parts, the high-end dual-core Atom is more likely to capture headlines, but it’s the lower-end Z2000 that’s driving Intel’s grand vision. India and China are two of the fastest-growing cell phone markets on the planet. In these markets, the Z2000 will be going up against single-core phones based on the ARM11 architecture that pre-dates even the Cortex-A8. The Z2000′s capabilities will make it an extremely potent competitor for the likes of the HTC Explorer (pictured right) or Lenovo A60; both devices currently sell into the sub-$200 smartphone segment.

Intel may have chosen to focus on growing market share in China and India due in part to the difficulty of cracking the US. Here, large phone subsidies and Apple’s dominant market position make it difficult to tempt manufacturers into taking a chance on a new phone architecture. Apple’s lawsuits against Android handset manufacturers may not have directly impacted Intel’s push into the market, but it’s helped create a general culture of uncertainty. The confluence of these factors means it makes better sense for Intel to build momentum and perceived legitimacy outside the US. There will be Medfield handsets available to domestic customers, but we’re far from the only market.

We’re still months away from seeing Medfield phones shipping to US customers; Z2460 devices should start shipping in Q2/Q3, with Z2000 phones near the end of the year. The dual-core, 32nm Z2580 will ship to retail in early 2013, probably around the same time that Intel unveils its next-generation 22nm Atom device family. By that time, Qualcomm’s Krait family will be well-established, Cortex-A15 devices should also be starting to ship, and Nvidia’s updated Tegra 3 should be available. 2012 increasingly looks like it’ll be the calm before the storm — and Intel intends to be well dug in before the waves hit.

Updated @ 12:27: Intel has announced a partnership with India’s Lava to launch a 1.6GHz Medfield phone running Gingerbread 2.3. More details as they become available.

Source

How the world’s fastest supercomputer, Fujitsu’s K, saves on power and money

Tired of cell phones and Mobile World Congress? Grab a bit of supercomputing news here. Fujitsu shared some details on how it managed to keep its K supercomputer running so efficiently at ISSCC last week. The K is currently the highest-performing supercomputer in the world, with a maximum speed of 10.5PFLOPS and a total of 705,024 cores. Such systems are notoriously power hungry; power consumption is one of the major hurdles DARPA identified in its recent call to US researchers to reinvent computing.

K improves efficiency by fine-tuning the power supply voltage of each and every CPU. According to the researchers, this type of customization was necessary to deal with the rise in CPU manufacturing variation; such variances can lead to significant differences in power consumption and operating temperature. As a result, researchers were able to reduce CPU power consumption by an average of 7W per chip.

7W per core may not sound like much on a per-system basis, or even in a mid-sized server room. In the K’s case, such analysis knocked around one megawatt off the system’s total power consumption and reduced the annual operating cost by $1 million. The systems are also water cooled (water being substantially more efficient than air), which helps reduce costs and improve performance.

Expect to see more of this type of optimization in the future — a lot more. The work the K team did in hand-tuning power supply voltages per physical CPU is exactly the sort of technique we expect to see adopted, potentially at every level. One of the issues plaguing semiconductor design teams is that as process nodes shrink, manufacturing variances become increasingly problematic; a tiny number of sub-optimal variances can lead to dramatically higher power consumption if they occur in the wrong places. This is one issue that Moore’s law-style transistor density actually exacerbates; the more transistors per die, the greater the chance of flaws in the manufacturing process.

While DARPA will undoubtedly be interested in applying these lessons to its own exascale computing initiative, we’d be surprised if mobile manufacturers weren’t taking their own notes. With transistor power consumption bottoming out as voltages approach minimum thresholds, power savings through optimizing other parts of the delivery system will be a prime target and a means of product differentiation.

Source

Goliath wins: AMD retreats, retrenches, and seeks to reinvent itself

AMD’s Financial Analyst Day packed a great deal of information into a few hours. It was, at times, frustratingly vague as far as the company’s long-term strategic plans, but one fact came through clearly. The performance war between Intel and AMD is over. Intel won.

Whether or not this is a bad thing is very much open to discussion. Let’s have a look at the big picture.

CEO Rory Read has said for months that the company’s focus would be in “low power, emerging markets, and the cloud.” New details on that plan, and Read’s comments yesterday, indicate that AMD will no longer focus on delivering high-performance products built on the latest process technology. The following slides lay out the current plan and future focus in stark terms.

Missing from the second slide is any mention of enterprise products (apart from cloud servers), discrete graphics, or conventional desktops. That’s not to say that AMD plans to abandon these markets, but none of AMD’s executive presentations or breakout discussions discussed high-performance CPU or GPU products. Instead, the company’s focus is on moving into SoC (System on a Chip) designs and leveraging that IP to create new products both for the 2012-2013 timeframes we’ve seen and beyond.

This move actually makes a lot of sense. The trend towards SoCs is driven by continued growth intransistor densities. AMD wants to be very active in the low power market, and bringing components on-die reduces both power consumption and communication latencies. This dovetails with the company’s decision to abandon its competition with Intel at the upper end of the market in several ways. AMD has no prior experience in building SoCs (unless you count the failed Krishna/Wichita designs), and such parts are inherently more complex than CPUs or even APUs. Combining new architectures and new processes in 2011 got AMD a chip it couldn’t manufacture, a chip nobody wants, and an APU it had to cancel. The only APU to hit its targets was Brazos — and it was built on 40nm.

Foundry economics are likely another factor. The press pays a great deal of attention to TSMC and GlobalFoundries’s production at their leading nodes, but that’s only a small component of their business. TSMC’s figures for Q4 2011 indicate that nearly 75% of the company’s revenue is built on process technology that debuted five years ago. With the cost and difficulty of each node transition rising, the foundries have little incentive to push the envelope. It makes much more sense for TSMC and GlobalFoundries to ramp nodes slowly — and that doesn’t fit well into a narrative of going head to head with Intel. AMD, in turn, isn’t willing to pay the hefty premium associated with driving early risk production at either foundry.

Ambidextrous AMD

AMD made multiple comments yesterday about being “flexible around ISA” and portraying itself as a “solutions” provider rather than being wedded to a single type of solution. At the same time, the company’s roadmaps through 2013 demonstrate a commitment to x86-based products.

Even if AMD is off designing an ARM core, it’ll take several years to bring a solution to market. Sunnyvale’s remarks around ISA (instruction set architecture) are best understood as part of the company’s strategic shift towards SoC’s and the mobile market.

What AMD is saying here is that if there is third-party IP it can license to improve its products, it’ll license it. If it sees an opportunity to develop its own IP for cross-licensing, it’ll do so. Part of the “flexibility” the company is discussing refers to HSA (Heterogeneous System Architecture) and the use of the GPU to handle specific tasks. The company is angling for shorter design cycles and a faster time to market, with new products shipping in 18-24 months as opposed to the 3+ years between CPU architectures. Again, the goal is to re-use synthesizable IP blocks rather than to create new products from the ground up.

HSA: The last bright spot

Nearly all of AMD’s presentations yesterday revolved around the need to reposition the company, build SoCs, and transition away from challenging Intel at the high end of the market. The one exception to that trend was AMD’s discussion of heterogeneous computing. Here, the company’s unification plans for CPU and GPU could still produce exciting products. AMD’s goal for HSA is to create solutions that are easier to program, optimize, and load balance while offering higher performance in a lower power envelope.

Here is AMD’s HSA roadmap. In 2012, we gain user mode scheduling and bi-directional power management. User mode scheduling is a Windows feature that allows applications to schedule their own threads without involving the system scheduler. Microsoft recommends it “for applications with high performance requirements that need to efficiently run many threads concurrently on multiprocessor or multicore systems.” AMD’s slide mentions C++ support for GPU compute, but that’s a feature of the HD 7900 family — APUs will have to wait for 2013.

The features scheduled for 2013 make it easier for the CPU and GPU to share data. A fully coherent memory model will allow the CPU to spin threads off to the GPU (or vice versa); the unified address space and use of pageable system memory simplifies communication between the two. By 2014, the boundaries between a “CPU workload” and a “GPU workload” have essentially become invisible. The idea is that a developer can target the HSA model and leave the decision of where code should be running up to the hardware.

HSA isn’t a replacement to OpenCL — AMD refers to it as an “optimized platform architecture” that uses OpenCL. The HSA ISA is virtual, meaning that it’s not tied to x86, ARM, or any particular language. According to Joe Macri, CTO of AMD’s client division, code written in HSA is compiled for the underlying CPU or GPU using a JIT (just in time) compiler.

This isn’t the first time we’ve seen this information from AMD, but the company never put dates on anything before. If AMD can build support for HSA, it could give its consumer products a substantial visibility kick.

End of an era

The presentations and speeches yesterday made it clear that the old AMD — specifically, the AMD that prioritized competing head-to-head with Intel and building products that could challenge the best Chipzilla had to offer — is dead. As an enthusiast who bought his first K6-233 in 1997, that’s hard to write. I cut my overclocking teeth on a Duron 600 with a pencil modded voltage of 1.85v that enabled the chip to boot at 160MHz on a KT-133A-based IWILL KK266; I still own the Tonicom PC166 RAM I bought for the system.

A moment of silence, if you please.

As a journalist who’s covered the company for a decade, however, it’s hard to ignore the fact that AMD isn’t giving up so much as it’s acknowledging that it lost this fight years ago. The company isn’t folding — it’s retrenching around the idea of building mainstream products for the mass market using proven process nodes. As strategies go, it’s still a risky one. Sunnyvale’s decision to focus on SoC integration makes sense given where the market is headed, but its 28nm Brazos will have to compete against 22nm Atom and 28nm ARM chips. If Piledriver performs well and AMD can ship its 2013 products towards the front of the year, it should do well. If its 2013 roadmap piles up in the back half, it risks being left behind altogether.

To be honest, Read’s decision to wrap up his presentation with this slide doesn’t engender confidence. What comes after “Shaping the future” — “Terrorize Tokyo with my radioactive breath weapon?”

AMD could still have a role to play in driving the evolution of mobile products and helping keep prices low by providing competitive performance in mainstream and value market segments. If the company’s HSA initiative takes off, it might even give Sunnyvale a long term launch point back into the higher-end desktop, mobile, and server markets. For now, however, the company is retreating from these markets — and leaving Intel Inside.

For further reading check out AMD’s Analyst Day confirms APU cancellations, trims core counts.

Source

The state of Chrome OS

A look at the Chrome Web Store shows a vibrant, lively catalog of software, allowing you to do anything from photo editing to playing games right in your browser. It is populated with many of the same titles as you would see in Apple’s iTunes App Store, from well-known applications to popular games.

Is it enough, though? Chrome OS has an edge over conventional operating systems — it’s only a web browser, after all. Low power consumption, low spec-requirements, and fast, reliable ease of use are its calling cards and it does those things extremely well… for a cost. Chrome OS doesn’t allow you to install actual applications at all; you are going to the Chrome Web Store in order to download (or purchase) web applications, which are really just shortcuts that appear on your “desktop” when you open a new tab.

Google Chrome OS was initially announced in July of 2009 and made open source as the Chromium Project that November, though the first beta test Cr-48 laptops (Chromebooks) didn’t ship until late December 2010. The Chrome Web Store has been filled to bursting with applications since that time. When Chrome OS was new and fledgling, many naysayers claimed it would not be robust enough, and that the breadth of web applications was too slim to really make a splash in the PC pond.

They could not be more wrong.

If you’re a power user that requires heavy computing power for resource intensive applications like video editing, CAD drawing and the like, Chrome OS is probably not going to be what you need – yet. The wonders of HTML5, Canvas, and enhancements to back-end web coding may one day make the need for beefy desktops go the way of the dinosaurs.

For the casual user, Chrome OS presents an opportunity to put all of your eggs into one basket. You turn it on (and it’s on fast) and you’re almost immediately looking at the icons of your applications after you log in. Your most-visited sites and recently-closed tabs are located conveniently at the bottom of the screen.

It can be pretty safely assumed that the average user will probably browse some news sites, play on social networks, perhaps watch a streaming video or two, and check their finances, perhaps in an afternoon at the PC. Chrome OS excels at all of those things. Users a bit more savvy might want to upload pictures to the web, and may need to edit those pictures before they’re ready to be seen in Google Plus; Chrome OS has those users covered as well. With themes and extensions, Google has made a simple browser operating system as customizable as it can be.

Speaking of a simple browser operating system, it should be said that saying that isn’t exactly true. Chrome OS is actually a web browser built on top of a custom Linux-based operating system. The user is kept in a sandbox, not allowed to install applications directly or change system settings. This creates a very secure environment for everyday computing without the need for antimalware or antivirus scanners.

Let’s take a casual user’s day and replace common applications with ones easily found (for free!) in the Web Store. Jane, let’s say, wakes up in the morning and checks her email first thing with a cup of coffee. After opening Gmail she browses her new messages, and then simply clicks on her Feedlyextension to instantly open a new tab containing all of the news from her Google Reader RSS feed, presented in a full, feature-rich magazine-like display that syncs with Jane’s social media as well.

Done with her morning coffee, Jane sits down to work. As a web designer, Jane needs a wide assortment of tools that are very specific to her job. She writes her code in Handcraft, Pendule oriScribe Tools, depending on what facet of coding she’s working on that day. All are WYSISWG editors, with iScribe having a feature-rich offline mode with code converter and code tidy applications. Image work is done in Pixlr Editor, a web-based editor so like Gimp or even Photoshop that you wouldn’t believe it’s free.

Occasionally Jane must do some of her own PHP or Ruby coding, and for that she uses Koding, a great online code editor with collaboration tools and a sync with SVN, Git, or Aviary, to keep her team in the loop. Speaking of her team, she’s in constant contact with them through Google Chat, and never misses an online meeting done through Gchat’s built-in video conferencing or the Hangout feature of Google Plus. She keeps track of her time and projects via TeamWork Live, and never has any issues knowing where she’s at with a particular project or schedule.

Every tool necessary is at her fingertips for a flowing, productive day, all in the browser.

What’s the catch? There’s always a catch, isn’t there? In Chrome OS’s case, its biggest asset is also its most glaring weakness: The internet. You have Google’s vision of the world at your fingertips when you’re connected. If you don’t have any internet access, though, there’s not much you can do with a browser OS that runs web applications. Your trusty Chrome OS laptop isn’t something you can take with you on that remote camping trip; although to be fair the Web Store has more and more offline tools popping up every day. There’s room to grow there, certainly, but users should be aware that Google’s OS brainchild is meant to be tethered to the internet in some fashion.

Now, power users do have some realistic issues to deal with. Chromebooks aren’t exactly brimming with power, so you’ll notice some heavy sites loading slowly as the hardware strains a bit. If you tend to have a ton of tabs open, you’ll notice everything slowing down the more you open, and there’s always slight latency in the machine when a new tab is loaded. These things probably wouldn’t be noticed by an average user, but a power user already skeptical of not being able to install their favorite apps will probably be looking for such difficulties, and they do exist.

Conclusion

What’s the bottom line? Chrome OS is a lively mobile operating system that contains more computing power than most users need. Its abundance of tools available online make it the go-to machine in almost every situation where portability, speed, and ease of use factor highly. For those that require more computing power there’s always Chromium, Google’s open source of Chrome OS that can be installed on almost any hardware, affording users more punch for their apps.

As a writer, I almost exclusively use my now-long-in-the-tooth Cr-48 Chrome OS laptop I received back in December of 2010 to sit and write with. Some of the newer Web Store apps are starting to give it a hard time, as the newer Chromebook hardware has improvements and upgrades my little laptop lacks. I’m a power user in almost every sense of the word, but with the portability and comfort that Chrome OS affords me I’m willing to sacrifice a bit of speed and raw computing power. I rarely leave home without it, and it’s the only computer I bring to events and trips.

It might have been released ahead of its time, in a world still gripped tight in the meaty fist of Microsoft’s Windows empire, but Chrome OS is a star that’s rising. I’d keep an eye on it, and if you get a chance, try it out. You won’t regret it.

Source

Microsoft patches Bulldozer’s performance — we investigate

When AMD launched Bulldozer back in October, the company claimed that problems with Windows 7 thread scheduler prevented the CPU from delivering peak performance. According to AMD, Windows 7 didn’t accurately understand how best to schedule threads to take advantage of the company’s shared architecture, and as a result “there are possibilities where opportunities for resource sharing or activate [sic] higher Turbo Core frequencies are missed.”

Microsoft has just released a pair of hotfixes that claim to resolve the issues that handicapped Bulldozer. Curious to see what might have changed, we dusted off our test rig, installed both updates, and ran some of the same tests we’d previously checked.

The updates in question are KB2646060 and KB2645594. The first changes how often Bulldozer achieves the C6 power state. MS notes that “this potentially results in increased power consumption in more lightly-threaded environments.” The second update is more vague, mentioning that “multithreaded workloads may not be optimally distributed… in a lightly-threaded environment. This may result in decreased system performance for some applications.”

AMD’s own statements on how much the update improves performance are similarly modest. “Our testing shows that not every application realizes a performance boost. In fact, heavily threaded apps (those designed to use all 8 cores), get little or no uplift from this hotfix – they are already maxing out the processor. In other cases, the uplift averages out to a 1-2 percent uplift.” writes Adam Kozak, AMD’s product marketing manager.

Here’s what we saw when we installed the updates and started testing. As in our original article, Turbo Core was disabled. A 4M/8C configuration means that a test was run on all four modules with both cores enabled per module. Previously, a 4M/4C configuration was notably faster than 4M/8C in quad-threaded workloads do to scheduling problems.

The good news is that these updates improve Bulldozer’s performance by 2.74% in Cinebench and 3.33% in Maxwell Render. The bad news is just how limited the updates are. As AMD notes, these boosts are limited to “lightly threaded environments” — we saw no performance improvement when we benchmarked these tests with all eight cores enabled.

Hopefully the updated scheduler and AMD’s own admission that it’s of limited use will help put a finish to persistent rumors that reviewers, Intel, and the Illuminati have somehow conspired to make Bulldozer’s performance out to be worse than it actually is. AMD’s demos at CES 2012 have focused nearly exclusively on the company’s graphics performance, as evidenced by this screenshot of the company’s CES page:

The demos of up-and-running Trinity silicon confirmed reports that the chip is ramping well and that GPU performance is a substantial improvement over Llano, but too many websites have combined two separate statements from AMD. First, that Trinity is based on Bulldozer and has up to four cores, and second, that AMD is targeting a 17W power envelope for its own ultra thin-and-light segment.

The fact that both of these statements are true doesn’t mean that AMD is planning to launch a 17W quad-core based on Piledriver. Even Intel never tried to hit anything near that with a mainstream Sandy Bridge part, opting instead to wait for 22nm and the debut of Ivy Bridge.

As we head towards AMD’s Financial Analyst Day in February, everything is rosy on the graphics side — at least, apart from the company’s inability to earn any significant profit from the business — but the CPU side of the equation is full of questions. Bulldozer isn’t getting any magic software-side improvements. Trinity’s GPU is awesomesauce, but the CPU will be doing well to match Llano’s performance-per-watt and 28nm follow-ups to Bobcat remain conspicuous in their absence.

Source