Be among the first to see Intel’s new Many Integrated Core Architecture in action. This is a short video of the demo that was shown at International Supercomputing conference in Hamburg Germany 2010. This 32 core – 128 thread co-processor is currently still in development. This software development kit is codenamed Knights Ferry and is the first product in a line of Knights family of coprocessors.

Click here to view the embedded video.

Fusion is a single, power-efficient chip that combines both the CPU and GPU into a single die – the APU. “Hundreds of millions of us now create, interact with, and share intensely visual digital content,” said Rick Bergman, senior vice president and general manager, AMD Product Group. “This explosion in multimedia requires new applications and new ways to manage and manipulate data. Low resolution video needs to be up-scaled for larger screens, HD video must be shrunk for smart phones, and home movies need to be stabilized and cleaned up for more enjoyable viewing. When AMD formally launches the AMD Fusion family of APUs, scheduled for the first half of in 2011, we expect the PC experience to evolve dramatically.”

In addition to the announcement and the demon of the new APUs, Rick Bergman also introduced AMD’s new Fusion Fund, a program that is intended to promote and support the development of new solutions that will take advantage of the processing power and design of AMD’s future Fusion APUs. Details about the new Fusion Fund are available at the company’s website.

“The computing industry is undergoing a long-envisioned and inevitable transition from text-based computing to a digital world of natural human-to-system interfaces and rich visual experiences that are incredibly vivid and realistic,” said Manju Hegde, corporate vice president, AMD Fusion Experience Program, and Fusion Fund managing partner. “AMD Fusion APUs are designed to expand the art of the possible for software developers, accelerate application performance, and deliver levels of power efficiency conducive to cooler and sleeker PC devices. The Fusion Fund will help expand the Fusion ecosystem focused on realizing the next generation of computing for the mainstream market.”

A video of AMD’s Fusion in action is available below, doesn’t look half bad for sure.

Click here to view the embedded video.

According to AMD, the Barzos platform that will be paired with the Fusion processors will be targeted at ultra-thin notebooks. They will supposedly operate on less power than current solutions, which means that the chip and platform might be very suited for tablet PCs. This fact comes even though AMD is “currently not painting a strategy” around such devices. The Fusion APU (accelerated processing unit) will be codenamed Ontario and will integrate the CPU and GPU on a single die, similar to what Intel’s 32nm CPUs do.

All APUs will be made based on the 32nm manufacturing process. They will target not just ultra-thin and ultra-mobile systems, but the line will also feature desktop PC offerings. The company will still monitor the “exciting” tablet market and will watch out for an opportune moment to enter it.

Most speculations around the APU deal with the actual graphics capabilities. Considering the extensive graphical know-how from ATI, there is little doubt that AMD should be capable of making APUs capable of running even the most modern games, even with all the difficulties of combining the two technologies into a single die. However, making such a super powered chip might actually threaten AMD’s graphics card sales. As such, while the Fusion is expected to surpass Intel’s 32nm chips, a high-end graphics performance may be out of the question.

AMD’s soon to be launched Netbook platform, Congo, is now renamed to simply “2nd Gen UT”. With 2010, we have the Danube platform, with AMD’s first notebook quad-core CPU. In terms of GPU, 2009 is listed as ATI M9x, which is Mobility HD 4000 series. Once again, it is listed as 40nm, though the only 40nm GPU is RV740

These happen to be famous avenues in Manhattan, so the Manhattan family codename fits in well. However, notice the fab process – 32nm! Manhattan was widely rumored to be straight derivates of desktop 40nm Evergreen GPUs. Even more surprising is that it appears the DX10.1 from M9x carries on into Manhattan. Barring typos and confusing slide designs, these are both major surprises. It would appear Manhattan is not a 40nm Evergreen based GPU, but rather a 32nm M9x shrink, something that seems very unlikely. We will look out for clarification on this information.

Moving on to 2011, and Fusion hits notebooks. Sabine platform is based around the same Llano APU found in desktops. Brazos is the low-voltage platform, which seems to be manufactured at 40nm – for the complete APU.

GPU for notebooks remains next-gen Northern Islands at 32nm. The 2011 platforms increases battery life (battery life being a constant criticism for AMD’s notebook platforms) by a significant amount [via vrzone].

iCooler x4 features a high performance dual slot cooler. It comes with four 6mm copper heatpipes that transfer heat effectively from the GPU to the internal fins around.

Unfortunately the clock frequencies were left alone as the card is clocked at 850 MHz on the core and has a memory clock 3.9Gbps / MHz. Pricing remains TBA.

Mind you that this will be the same for any DX11 class graphics adapter, and maybe even some DX 10 class graphics cards

Check out the pics.

That is until now, TechConnect Magazine has some great images of the low-end 40nm desktop card from nVidia. The new G210 and GT 220 will feature some new stuff (besides the 40nm shrink). The G210 will have 16 sores, a 64-Bit memory interface, DX 10.1 support and 512MB of memory. The GT 220 will have 48 Cores, 128-Bit memory interface and double the memory.

While pricing is not known they should hit the shelves sometime in Q3, while some OEMs will have access to them sooner.

I am not talking about chipsets either. It seems that Intel might have got eVGA all excited about the Larrabee GPU and they are looking to push them on the public.

As it stands eVGA is one of the last nVidia only AIBs out. They have increased their reputation by standing behind their cards with Life-Time warranties, Upgrade programs, and free overclocking tools and software (might explain why they do not work with AMD/ATi).

If this is the case and eVGA moves to Larrabee it will mean that the first Boards with this new GPU will at least come from a good company.

EXPECTATION DICTATES that Intel’s long-awaited Larrabee offerings will follow the same production model pioneered by rival Nvidia at the start of the decade. But what the Green Goblin might not have counted on is that its most loyal partner would be leading the charge to bring Chipzilla’s products to market.

By all accounts, EVGA is currently Nvidia’s number one partner, having shown 100 per cent loyalty to Jen Hsun Huang and his Green Goblin… until now.

Following a path blazed by Asus, Gigabyte and MSI, EVGA looks set to be the latest to jump the Nvidian ship and seek other product partners.

Intel is current market leader with 49.7% and 37.20 million sold untis, Nvidia is second with 31.1% market share and 23.26 million untis, and AMD/ATI was only able to sell 12.81 million units, bringing their market share down to 17.1%. Another interesting observation was that after eight consecutive years, for the first time, the negative sales from Q4 to Q1 turned into positive sales, with a 3.29% change from Q4 2008 to Q1 2009.

All in all and compared year to year, things didn’t look so good, though – which had to be expected, considering the current economical crisis. In raw numbers this means that “shipments dropped to 74.9 million units, -21.1% from this quarter last year.”

Jon Peddie Research estimates that “the cleaning out of old inventory, both in the electronics, automotive, and real-estate markets should be pretty well completed by the end of the summer (and some amazing deals will be available if you are happy with 2008 technology).”

Next to that, NVIDIA supposedly is moving from SIMD to MIMD architecture. Long story short: SIMD always has the advantage in raw peak operations per second. After all, it mainly consists of as many adders, floating-point units, shaders.

SIMD (Single Instruction, Multiple Data) refers to a parallel computer that runs the exact same program, that’s the “single instruction” part – on each of its simultaneously-executing parallel units. MIMD can pack more of their stuff onto chips, too, using the same density.

MIMD (Multiple Instruction, Multiple Data) refers to parallel computing that runs an independent separate program – that’s the “multiple instruction” part – on each of its simultaneously-executing parallel units.

Why NVIDIA makes this move is a little unclear, but basically the GPU becomes a huge floating point unit, which is great for games of course .. yet even better for GPGPU activity and CUDA related applications.

The new GPU will have 16 clusters of 32 shaders. Though it is unknown how fast they will be clocked, rumors state that if they can be clocked at 2 GHz, we’d be nearing a cool 3 teraflops of compute performance (right now we are hovering around one teraflop) per GPU. GT300 would give direct hardware access [HAL] for CUDA 3.0, DirectX 11, OpenGL 3.1 and OpenCL.

There is no release date on GT300 but we do know it’ll be somewhere this year. But from the looks of it, that’s a big can of you know what that NVIDIA will be delivering: