Advantech, a leading provider of AIoT platforms and services, proudly unveils its latest AI acceleration modules: the EAI-1200 and EAI-3300, powered by Hailo-8 AI processors. These modules deliver AI performance of up to 52 TOPS while achieving more than 12 times the power efficiency of comparable AI modules and GPU cards. Designed in standard M.2 and PCIe form factors, the EAI-1200 and EAI-3300 can be seamlessly integrated with diverse x86 and Arm-based platforms, enabling quick upgrades of existing systems and boards to incorporate AI capabilities. With these AI acceleration modules, developers can run inference efficiently on the Hailo-8 NPU while handling application processing primarily on the CPU, optimizing resource allocation. The modules are paired with user-friendly software toolkits, including the Edge AI SDK for seamless integration with HailoRT, the Dataflow Compiler for converting existing models, and TAPPAS, which offers pre-trained application examples. These features accelerate the development of edge-based vision AI applications.

EAI-1200 M.2 AI Module: Accelerating Development for Vision AI Security
The EAI-1200 is an M.2 AI module powered by a single Hailo-8 VPU, delivering up to 26 TOPS of computing performance while consuming approximately 5 watts of power. An optional heatsink supports operation in temperatures ranging from -40 to 65°C, ensuring easy integration. This cost-effective module is especially designed to bundle with Advantech's systems and boards, such as the ARK-1221L, AIR-150, and AFE-R770, enhancing AI applications including baggage screening, workforce safety, and autonomous mobile robots (AMR).

A few days ago, we published a report about Intel's 18A yields being at an abysmal 10%. This sparked quite a lot of discussion among the tech community, as well as responses from industry analysts and Intel's now ex-CEO Pat Gelsinger. Today, we are diving into known information about Intel's 18A node and checking out what the yields of possible products could be, using tools such as Die Yield Calculator from SemiAnalysis. First, we know that the defect rate of the 18A node is 0.4 defects per cm². This information is from August, and up-to-date defect rates could be much lower, especially since semiconductor nodes tend to evolve even when they are production-ready. To measure yields, manufacturers use various yield models based on the information they have, like the aforementioned 0.4 defect density. Expressed in defects per square centimeter (def/cm²), it measures manufacturing process quality by quantifying the average number of defects present in each unit area of a semiconductor wafer.

Measuring yields is a complex task. Manufacturers design some smaller chips for mobile and some bigger chips for HPC tasks. Thus, these two would have different yields, as bigger chips require more silicon area and are more prone to defects. Smaller mobile chips occupy less silicon area, and defects occurring on the wafer often yield more usable chips than wasted silicon. Stating that a node only yields x% of usable chips is only one side of the story, as the size of the test production chip is not known. For example, NVIDIA's H100 die is measuring at 814 mm²—a size that is pushing modern manufacturing to its limits. The size of a modern photomask, the actual pattern mask used in printing the design of a chip to silicon wafer, is only 858 mm² (26x33 mm). Thus, that is the limit before exceeding the mask and needing a redesign. At that size, nodes are yielding much less usable chips than something like a 100 mm² mobile chip, where defects don't wreak havoc on the yield curve.

The Pokémon Trading Card Game Pocket is billed as a "casual," accessible way for people to enjoy the fun of the Pokémon Trading Card Game, regardless of where they are in the world. That said, it looks like players will also have to endure the added fun of the financial burden of the real-world Pokémon TCG if they want to be able to play the game in any serious fashion. According to one gamer's analysis posted on Reddit, however, the amount of time you'd need to invest into Pokémon Trading Card Game Pocket is anything but casual, clocking in at over two years of daily gameplay to unlock your first whole set.

The method used in the analysis was based on the likelihood of obtaining a new card needed to complete a card set in one of the two daily card draws available to non-premium players. Without optimizing gameplay using features, like the Points Shop or the Wonder Pick, both of which greatly increase the odds of drawing specific cards, players would need to play consecutively for 1843 days, opening a total of 3,687. Paying for a premium pass enables opening three packs a day, however, reducing that number to 1,229 days and resulting in a total bill of $404 for that period. It's only by making full use of the Wonder Pick and Points Shop that the total play time comes down to a little under two years, at a total of 655 days (around 1,310 packs). Upgrading to the premium pass reduces this, but will run up a total of $180 but reduce the amount of time significantly.

Intel has officially started shipping the "0x129" microcode update for its 13th and 14th generation "Raptor Lake" and "Raptor Lake Refresh" processors. This critical update is currently being pushed to all OEM/ODM partners to address the stability issues that Intel's processors have been facing. According to Intel, this microcode update fixes "incorrect voltage requests to the processor that are causing elevated operating voltage." Intel's analysis shows that the root cause of stability problems is caused by too high voltage during operation of the processor. These increases to voltage cause degradation that increases the minimum voltage required for stable operation. Intel calls this "Vmin"—it's a theoretical construct, not an actual voltage, think "speed for an airplane required to fly". The latest 0x129 microcode patch will limit the processor's voltage to no higher than 1.55 V, which should avoid further degradation. Overclocking is still supported, enthusiasts will have to disable the eTVB setting in their BIOS to push the processor beyond the 1.55 V threshold. The company's internal testing shows that the new default settings with limited voltages with standard run-to-run variations show minimal performance impact, with only a single game (Hitman 3: Dartmoor) showing degradation. For a full statement from Intel, see the quote below.

Dylan Patel, of SemiAnalysis, has highlighted worrying industry trends from an October 2021 published report—the Center for Security and Emerging Technology (CSET) document explored and "(outlined) infrastructure investments and regulatory reforms that could make the United States a more attractive place to build new chipmaking capacity and ensure continued U.S. access to key inputs for semiconductor manufacturing." Citing CSET/World Fab Forecast findings, Patel expressed his dissatisfaction with the apparent lack of progress in the region: "The United States is the slowest relevant country in the world to build a fab thanks to NIMBY assholes and the garbage regulatory/permitting system." The SemiAnalysis staffer likely believes that unsuitable conditions remain in place, and continue to hinder any forward momentum—for greenfield fabrications projects, at least.

The CSET 2021 report posited that the proposed $52 billion CHIPS Act fund would not solve all USA chip industry problems—throwing a large sum of money into the pot is not always a surefire solution: "The United States' ability to expeditiously construct fabs has declined at the same time as the total number of fab projects in the United States has declined. Some of this is due to changes in the global semiconductor value chain, which has concentrated resources in Asia as foundries have risen in prominence, and countries like Taiwan, South Korea, and China have established significant market share in the industry from 1990 to 2020. However, during this same 30-year period, the time required to build a new fab in the United States increased 38 percent, rising from an average of 665 days (1.8 years) during the 1990 to 2000 time period to 918 days (2.5 years) during the 2010-2020 time period (Figure 2). At the same time, the total number of new fab projects in the United States was halved, decreasing from 55 greenfield fab projects in the 1990-2000 time period to 22 greenfield fab projects between 2010 and 2020." Intel's work-in-progress Ohio fabrication site has suffered numerous setbacks (including delayed CHIPS Act payments)—the latest news articles suggest that an opening ceremony could occur in late 2026 or early 2027. Reportedly, TSMC's Arizona facility is a frequently runs into bureaucratic and logistical headaches—putting pressure on company leadership at their Hsinchu (Taiwan) headquarters.

DigiTimes Asia has reached out to insiders at fabrication toolmakers in an effort to delve deeper into claims made by industry analysts at the start of 2024—both SemiAnalysis and China Renaissance have proposed that TSMC is unlikely to adopt High-NA EUV production techniques within a five year period. The latest news article explores a non-upgrade approach for the next couple of years: "TSMC has not placed orders for high-numerical aperture (High-NA) extreme ultraviolet (EUV) tools and is unlikely to use the technology in 2 nm and 1.4 nm (A14) process manufacturing." Intel Foundry Services (IFS) will be one of the first semiconductor manufacturers to go online with ASML's latest and greatest machinery, although no firm timeframes have been confirmed. Team Blue's Taiwanese rival (and occasional business partner) is seemingly happy with its existing infrastructure, but industry watchdogs propose that cost considerations are key factors behind TSMC's cautious planning for the next decade.

The DigiTimes insider sources believe that TSMC will not budge until at least 2029, possibly coinciding with a 1 nm production node—analysts at China Renaissance reckon that High-NA EUV machines could be delivered in the future when facilities are readied for an "A10" codenamed process. TSMC published a very ambitious "transistor count" product timeline in early January (see below)—the first "1 nm" products are supposedly targeted for a 2030 rollout, but this schedule could change due to unforeseen circumstances. Intel is expected to "phase in" its fanciest ASML gear collection once the 18A process becomes old hat—Tom's Hardware thinks that 2026 - 2027 is a feasible timeframe.

Worldwide semiconductor revenue in 2023 totaled $533 billion, a decrease of 11.1% from 2022, according to preliminary results by Gartner, Inc.

"While the cyclicality in the semiconductor industry was present again in 2023, the market suffered a difficult year with memory revenue recording one of its worst declines in history," said Alan Priestley, VP Analyst at Gartner. "The underperforming market also negatively impacted several semiconductor vendors. Only 9 of the top 25 semiconductor vendors posted revenue growth in 2023, with 10 experiencing double-digit declines."

The combined semiconductor revenue of the top 25 semiconductor vendors declined 14.1% in 2023, accounting for 74.4% of the market, down from 77.2% in 2022.

The latest Canalys data reveals that PC shipments (desktops, notebooks and workstations) in the US declined just 6% year-on-year to 18.2 million units in Q2 2023, marking a significant improvement compared to earlier quarters this year. Notebook (including mobile workstations) shipments were down 4% to 15.2 million units, bolstered by the return of Chromebook demand in the education sector. Desktops (including desktop workstations) suffered a steeper decline, with shipments falling 12% to 3.0 million units. The US tablet market faced a similarly modest decline, with shipments down 5% to 10.3 million units.

"Despite undergoing another year-on-year decline, the US PC market showed promising signs of improvement in the second quarter," said Ishan Dutt, Principal Analyst at Canalys. "With the buildup of channel inventories now largely cleared, pockets of demand strength are now being reflected in vendors' sell-in shipment performance. A key area that helped drive volumes was the return of demand from education institutions, backed by the latest wave of federal funding, ahead of a licensing cost increase for ChromeOS. This helped propel Chromebook shipments to 4.7 million units, the highest level since peak deployments during the first half of 2021."

With sequential IC sales declines beginning to moderate, the global semiconductor industry appears to be nearing the end of a downcycle and is expected to begin to recover in 2024, SEMI, in partnership with TechInsights, reported in the Semiconductor Manufacturing Monitor. In Q3 2023, electronics sales are projected to post healthy quarter-on-quarter growth of 10%, while memory IC sales are expected to log double-digit growth for the first time since the downturn started in Q3 2022. Logic IC sales are predicted to remain stable and improve as demand gradually recovers.

Headwinds will continue for the semiconductor manufacturing sector in the second half of the year, SEMI and TechInsights reported. Drawdowns of high inventory at integrated device manufacturer (IDM) and fabless companies will continue to suppress fab utilization rates to much lower levels than those in the first half of 2023. The weakness is projected to extend declines in capital equipment billings and silicon shipments for the rest of the year despite stable results in the first half of 2023.

Valve's Steamworks Development team has reaffirmed, in an announcement earlier on this week, that the company is committed to protecting the privacy of its userbase: "Steam will continue to not share personally identifiable information. This approach to privacy means that some trade-offs have been made along the way that limits how specific some reporting can be. In most cases, it simply means that any traffic sources that are below a threshold of volume will get reported as "other". We intentionally don't collect or store demographic information about users such as age, gender, or race."

Valve has decided to drop its usage of Google Analytics as part of Steam's traffic monitoring package: "As of this coming July, Google will no longer operate Universal Analytics (UA), which is a third-party traffic reporting system that we've had available for measuring traffic sources to Steam store pages. Their announced replacement will be a system called Google Analytics 4 (GA4). You can read Google's help article about the topic." Valve has chosen to develop their own system instead of upgrading to the latest version of third-party software: "As time has gone on we've come to realize that Google's tracking solutions don't align well with our approach to customer privacy, and so with the migration to GA4 we've made the decision to end our support of Google's analytics systems on Steam. Instead, we're focused on building the most useful parts of aggregated reporting into Steam itself, as described above."

Over the past few weeks, the legal dispute between Arm Ltd. and Qualcomm Inc. has been warming up the eyes of the entire tech community. However, as per the latest court filing, Arm could change its licensing strategy and shift its whole business model into a new direction that would benefit the company directly. Currently, the company provides the intellectual property (IP) that chip makers can use and add to designs mixed with other IPs and custom in-house solutions. That is how the world of electronics design (EDA) works and how many companies operate. However, in the Qualcomm-Arm legal battle, Qualcomm's counterclaim has brought new light about Arm's plans for licensing its hardware designs past 2024.

According to Dylan Patel of SemiAnalysis, who examined court documents, Arm will reportedly change terms to use its IP where the use of other IP mixed with Arm IP is prohibited. If a chip maker plans to use Arm CPU IP, they must also use Arm's GPU/NPU/ISP/DSP IPs. This would result in devices that utilize every design the UK-based designer has to offer, and other IP makers will have to exclude their designs from the SoC. By doing this, Arm directly stands against deals like the Samsung-AMD deal, where AMD provides RDNA GPU IP and would force Samsung to use Arm's Mali GPU IP instead. This change should take effect in 2025 when every new license agreement has to comply with new rules.

Diversity is the name of the game when it comes to the edge Artificial Intelligence (AI) chipset industry. In 2019, the AI industry is witnessing the continual migration of AI workloads, particularly AI inference, to edge devices, including on-premise servers, gateways, and end-devices and sensors. Based on the AI development in 17 vertical markets, ABI Research, a global tech market advisory firm, estimates that the edge AI chipset market will grow from US $2.6 billion in 2019 to US $7.6 billion by 2024, with no vendor commanding more than 40% of the market.

The frontrunner of this market is NVIDIA, with a 39% revenue share in the first half of 2019. The GPU vendor has a strong presence in key AI verticals that are currently leading in AI deployments, such as automotive, camera systems, robotics, and smart manufacturing. "In the face of different use cases, NVIDIA chooses to release GPU chipsets with different computational and power budgets. In combination with its large developer ecosystem and partnerships with academic and research institutions, the chipset vendor has developed a strong foothold in the edge AI industry," said Lian Jye Su, Principal Analyst at ABI Research.

NVIDIA is facing stiff competition from Intel with its comprehensive chipset portfolio, from Xeon CPU to Mobileye and Movidius Myriad. At the same time, FPGA vendors, such as Xilinx, QuickLogic, and Lattice Semiconductor, are creating compelling solutions for industrial AI applications. One missing vertical from NVIDIA's wide footprint is consumer electronics, specifically smartphones. In recent years, AI processing in smartphones has been driven by smartphone chipset manufacturers and smartphone vendors, such as Qualcomm, Huawei, and Apple. In smart home applications, MediaTek and Amlogic are making their presence known through the widespread adoption of voice control front ends and smart appliances.

Prices of NAND flash could drop to historic lows of $0.08 per gigabyte in 2019, according to Jim Handy from Objective Analysis, addressing delegates at the 2018 Flash Memory Summit. If you add the cost of the controller, optional DRAM chip, and other low-cost parts that make up an SSD, 480~512 GB drives under $70 could finally be a reality; followed by 1 TB under $120, and 2 TB under $200. Handy attributes the low prices to a catastrophic oversupply of NAND flash in the industry, which could push manufacturers to the brink of economic collapse.

The price drop is also accelerated with the introduction of the QLC (4 bits per cell) technology, which increases densities (and conversely decreases price/GB). Luckily, most NAND flash manufacturers also happen to make DRAM, and are offsetting some of their NAND flash losses with DRAM profits, as DRAM remains in undersupply. The NAND flash price-crash threatens to wipe out conventional hard-disk drives from the consumer-space, at least in matured markets; relegating them to developing markets.

According to a new market report published by Transparency Market Research "Solid State Drive Market - Global Industry Analysis, Size, Share, Growth, Trends and Forecast 2015 - 2022," the global Solid State Drive (SSD) market was valued at US$15.1 billion in 2014 and is expected to reach US$229.5 billion by 2022, expanding at a CAGR of 40.7% from 2015 to 2022.

Asia Pacific was the largest revenue contributor in 2014 and accounted for 57.0% of the global solid state drive market. The higher performance requirement in enterprise applications and consumer mobile devices is fueling the demand for solid state drives. Features such as lower power consumption, cooler computing, low form factor, and reduced latency are cementing the demand for solid state drives. The solid state drive market has witnessed rising demand for client SSDs, owing to their use in devices such as notebooks, tablets, and mobiles. Asia Pacific is expected to be the fastest-growing market for solid state drive in terms of revenue during the forecast period. Growth in this region is mainly attributed to an expanding customer base and high IT infrastructure spending in countries such as Japan, China, and Taiwan. Competition in the market is paving way for the development of novel technologies in the solid state drive market. Furthermore, benefits offered by SSDs over HDDs, and the resulting replacement demand, are expected to drive the growth of solid state drive market during the forecast period.

TRENDFOCUS recently published its updated Quarterly NAND/SSD report covering the NAND & SSD industries' performance in CQ2 '15. In the SSD market, unit shipments increased, despite the fact that a good portion of the volume is tied to a weak notebook PC market. The two main factors contributing to the slight 2.9% Q-Q rise was the increase in DFF (drive form factor) client SSDs in the channel market, as well as an uptick in datacenter demand for SATA SSDs, driving a total of 23.859 million units. Total SSD capacity shipped increased by 13.3% Q-Q to 6.4 exabytes, with the overall average capacity increasing to 268 GB. Despite the rise in overall unit shipments, client modules declined, reflecting the continued slow demand for notebook PCs.

Enterprise SAS SSDs declined 10% while PCIe SSDs in the enterprise segment remained under 100,000 units for the quarter. Enterprise SATA SSDs were the bright spot in the enterprise market, posting an increase of 48.6% Q-Q as datacenters resumed purchases of SSDs - a trend not seen in nearline HDDs during the quarter. Samsung continues to dominate due to its successes in the client space - both from a unit and exabyte perspective - while Kingston showed once again that it can be a major player by focusing on the channel market. HGST's unit and exabyte share leads the SAS SSD market and Intel continues to dominate the enterprise PCIe market for a second quarter in a row.

AMD today announced the availability of GPU PerfStudio 2.0, a free, platform-agnostic, next-generation AMD graphics application development tool for game developers, obtainable through a closed beta program. GPU PerfStudio 2.0 has been designed from the ground up to easily integrate with existing projects, and help to identify and resolve performance issues early in the development cycle. With these tools, AMD is bringing its unparalleled understanding of 3D graphics technology and the leadership of its line of ATI Radeon graphics cards to bear on the development process.

"The GPU PerfStudio 2.0 suite of tools is extensive, customizable and, most importantly, based on open standards, which is critical for developers looking to integrate them into their existing workflows," said Neal Robison, director, Developer Relations, Graphics Products Group, AMD. "AMD is committed to arming developers with a full complement of hardware and development tools that help increase their development efficiency and improve 3D performance quality."

Today, Intel made its newest piece of silicon, the Core i7 series processor official with the introduction of three models based on the newer architecture, and socket. One of the significant feature-additions for Core i7 is the 192-bit wide DDR3 memory bus, meaning that memory modules in groups of three or six can be used to make use of the triple-channel memory controller the processors come with.

For memory vendors, it means selling kits consisting of three or six modules of 1 GB or 2 GB per module, resulting in 3 GB, 6 GB or potentially, a 12 GB kit (consisting of six 2 GB modules). Corsair, on its part, has been advertising the benefits of 4 GB system memory in the recent past, in a bit to sell its 2x 2GB kits. One of the important benefits Corsair noted was, that it benefited today's games. Come Core i7, and Corsair thinks 6 GB of memory would significantly benefit gaming PCs based on the Core i7 processors. The company released a whitepaper, in which gaming performance between Core i7-based systems equipped with 3 GB and 6 GB of memory were compared.

The market for laptops is quite uncertain at the moment. With Apple having made/making a transition over to Intel, ASUS launching more EeePCs and an anticipated flood of cheap computers, the market could go anywhere from its current state. Analysts over at Gartner have a pessimistic view of the future. They think that the very thing that got people into laptops in the first place could doom the industry. A desire for capable, portable information powerhouses could, by 2012, push businesses and people away from laptops, and towards smaller devices such as PDAs and so-called smart phones.

Manufacturers are nowadays bound to their respective marketing gurus resulting in ridiculously exaggerated trends which the usual customer tends to overlook. Let us talk about power supplies for instance. With 1000 Watt power supplies available from many companies and 2000 Watt power supplies just around the corner one can come quickly to the conclusion that you really need that much power.
But you don't! Mostly every website out there who puts PC hardware states the electric power consumption or gives a rough figure about it. If you then add those numbers together for every component you have you might be surprised about the result. For Example, the guys from Legion Hardware put together a dead fast machine consisting of a quad-core Intel CPU, two GeForce 8800GTX cards, a feature-packed ASUS Striker Extreme motherboard and four 1GB DDR2 memory modules. Additionally the system had four Seagate 320GB hard drives and six 120mm case fans and let's not forget the water-cooling system. Now reckon what wattage the power supply they used to power that beast was rated at!
550 Watt - They took a Thermaltake Purepower 550w and it works like a charm for more than two months now.

Power supply output ratings are actually quite inaccurate and can be very misleading. This is why it is very important to purchase a trusted and respected brand name.

If you came this far I would suggest you reading that article, it's well written and outlines every aspect of what to have in mind when purchasing a decent power supply.

Besides the usual R600 and K10/Barcelona talk that keeps us busy all day long there is another side of AMD that many people don't seem to realize: the financial part. If you are interested on being reading some well written thoughts I would suggest a recent two-part article at Overclockers.com. There Ed Stroligo explains AMDs actual financial situation with all the losses of revenue during the last quarter and gives some forecasts on how AMD will react to this.

Get to the first part here and don't miss the second part.

Btw.: There are several other interesting articles about various subjects of the computer sector at their front page.

Hardcore Windows XP fans will claim that Vista is currently buggy and bloated. Hardcore Mac OS X fans will claim that at least half of Vista is inspired by Mac OS X. Hardcore Linux fans laugh at all of this, and simply install the free Beryl user interface on the latest free distro of Ubuntu (or some other *nix distro). However, The Register has found some compelling reasons to move from whatever you're using to Windows Vista. The following is a shortened list, please click the source link for the full version of the list.

• UAC- It really makes the OS more secure, and despite how annoyed you may get with it, you have to admit it's a step in the right direction.
• Windows Aero- It's pretty and easy. Enough said.
• It comes with better bundled software- All the programs that come pre-loaded with Windows are either old favorites, or very new programs with great features.
• Vista Live = Xbox Live for the PC. Really.
• Halo 2- What Halo fans have been waiting for over the years is finally coming to the PC. And it looks better than most of us imagined.
• DirectX10 API- All new games/ graphics cards will eventually be exclusively for DX10 and Vista, so we better get used to it.
• Windows Vista search functions- Windows indexing the everything really helps when you misplace something.
• Windows Firewall- It's pretty good this time around, if you need a firewall.
• Lots of pretty hardware/laptops are coming out that are designed just for Vista.
• Microsoft is pushing for a complete move to Windows Vista, hence, they'll drop support for everything older just like they did for Windows 9x and ME.

AMD's ATI has a fancy name for their graphics card computing solution (Stream computing), so why doesn't NVIDIA? Since the CUDA name isn't exactly very appealing to the consumer market, NVIDIA is planning on renaming the CUDA incentive when it has something more ready for public release (such as a Folding@Home client). Rumor has it that NVIDIA will call their GPU-accelerated computing solution "GPU Computing", but we will see when NVIDIA actually releases clients to the average user.

For those of you that don't know, CUDA is a C compiler that compiles software so that it can be run/accelerated using an NVIDIA GPU as a processor.

People who want the right to rip their own DVD movies onto their computers have just won a major battle. The DVD Copy Control Association was trying to stop a California-based startup company from selling their only product, which ripped DVD's onto computers. When the DVD Copy Control Association brought this to court, the judge and jury took one look at the license and declared that not letting this company do their thing would violate the DVD Copy Control Association's own license. The judge says that "This (the license) is a product of a committee of lawyers". Basically, the license that tried to enforce copy-protection laws really just ended up shooting the copy-protectors in the foot (figuratively speaking, of course).

On February 16th of this year, NVIDIA went ahead and announced their new GPU computing client, CUDA. Over seven weeks later, we have yet to see so much as a beta of CUDA. And so, this means that NVIDIA, similar to the G80 Vista driver fiasco, has yet to deliver a product that they promised we'd see. The part about this that NVIDIA users probably hate the most is the lack of a GPU-based Folding@Home client. It seems that these days, everything except an NVIDIA system (even the PS3) can run a Folding@Home GPU client. This is an extreme opposite to ATI's GPU processing client, called "Stream", which has a large list of clients that can be used to accelerate programs using the GPU.

Desktops have been the dominant computer format ever since personal computers first took off, but a new report has suggested that laptop sales will overtake those of desktops by 2011. Analysts predict that the laptop market will grow by 16.1% each year until 2011, with just 3.8% predicted for desktops - during 2006, 140 million desktops PCs were shipped worldwide with 82 million portable computers being sold, up 2% and 26.3% respectively from 2005. Shifting towards mobile computing is becoming more popular as the performance gap gets smaller and notebook computers become cheaper, aided by the increased availability of wireless networks. In the USA, laptop computers have already begun outselling desktops, with 2005 being the first year in portable computers became the number one choice. Desktops are likely to remain the main choice for computer enthusiasts and those wanting to get maximum performance, as their easy upgradeability and more powerful components make them a more suitable option.