Galaxy have pleasantly surprised us, and the folks at vr-zone.  To the table, they bring their GTX 470 GC, a 100% non-reference design graphics card utilising an NVIDIA GeForce Fermi 470 GPU.  Measuring 9 inches (compared to the reference 9.5 inches) and featuring a blue PCB, Galaxy have essentially mated a graphics card with a robot figurine.

This 470 uses a quad heatpipe design, aluminium fins and a detachable fan (to help with cleaning) in order to cool the pre-overclocked behemoth; the core recieves a mild overclock of 18Mhz to 625Mhz, whereas shaders and memory stay at the reference 1250Mhz and 3348Mhz (effective) respectively.  This combines with the standard GTX 470 fare - DirectX 11, 448 CUDA Cores, 1280MB of GDDR5 memory, a 320-bit wide interface and 4x SLI compatible.

Check out the full story at VR-Zone.

The nForce 2 was one of the best chipsets to come out of NVIDIA. It was NVIDIA’s second attempt at a desktop chipset yet it cemented NVIDIA’s position as a leading provider of core logic in the market. Oh how much has changed since then.

Most of what made a chipset worthwhile is now integrated into the CPU. It used to be memory controller optimizations that kept chipsets afloat, but that’s no longer the case as all x86 processors now ship with an integrated memory controller. The performance differences between chipsets disappeared and all that was left was testing, validation and drivers, also known as the boring, expensive and time consuming aspects of chipset development.

NVIDIA has mostly abandoned the chipset business, focusing on opportunistic wins where possible. The most obvious example is with ION, its chipset/GPU for Atom based systems.

Last year we played with a number of ION systems and motherboards and were generally impressed. The platform made for a powerful little HTPC. Today we’re able to bring you a hands on performance preview of the first Next Generation ION box we’ve laid our hands on: the Zotac ZBOX HD-ID11.

That's Zotac's HD-ID11-U, a nettop based on the dual-core Atom D510 and NVIDIA's 16 core next-gen ION with 512MB DDR3 frame buffer. The system shipped entirely barebones so we'll be suiting it up and running whatever preliminary tests we can on it over the weekend. As I mentioned earlier, drivers aren't final so don't expect a full rundown anytime soon.

The system sports 6 USB ports, HDMI & DVI out, eSATA, Ethernet, optical audio out, headphone/mic jacks and a SD card reader. Internally you've got one 2.5" bay for a SATA HDD (or SSD) and a single SO-DIMM slot for DDR2.

If you want a closer look at the system head over to our pics in the Gallery.

Today, Sapphire have announced the latest in their lineup of ATI 5xxx series graphics cards - the passively cooled HD 5550, dubbed the 'Ultimate'.  

The party piece of this hardware comes in the form of the wrap around heatsink - the passive design and the low power usage (10W in 2D or idle mode) is destined to appeal to HTPC and quiet systems enthusiasts who yearn for DirectX 11.

Using a Sapphire-custom PCB and AMD's 40nm Redwood GPU, you will see 320 shaders clocked at 550Mhz, with a 1GB DDR2 memory at 800Mhz connected via a 128-bit interface.  Seems odd that they're using DDR2 and not GDDR3, given the increased heat dissipation mechanics of GDDR3 and the cost between them is negligible.

In terms of battling against the multimedia, the card offers DVI, HDMI and and VGA outputs while conforming to the HDMI 1.3a standard for full support on Dolby TrueHD and DTS-HD Master Audio.  ATI's on-board Unified Video Decoder will provide hardware decoding of all Hi-Def media.

Pricing so far has not been announced, but expect it to be in the region of £70/$90.

Brandon Hill, Editor in Chief of DailyTech, IMed me an hour ago with this: "OK, stop laying on the couch with your iPad and do some OS X benchmarking". He ended the IM with a link to a DT article stating that less than a week after Apple opened hooks into NVIDIA's VP2 decode engine, Adobe delivered a version of Flash 10.1 with GPU acceleration under OS X (Windows users have had it for six months now).

Impressive turnaround time for a company that has recently been thrashed by Apple quite a bit. It just goes to show one thing: there's no room for ego in engineering. Adobe claims the beta only supports Flash acceleration on the GeForce 9400M, GeForce 320M or GeForce GT 330M, however in my testing it worked fine on EVGA's GeForce GTX 285 Mac Edition. The tests below look at CPU utilization of the Flash plugin alone in Chrome (this is single core CPU utilization). The column on the left is without GPU acceleration, the one on the right with GPU acceleration:

While hardware acceleration doesn't appear to work on Hulu's website, there's definitely an improvement in CPU utilization when scaling to full screen. YouTube is a different story however. CPU utilization is cut roughly in half. The fact that it's taken this long is upsetting, but at least we're making some progress. You can tell the GPU acceleration is working if you see a little white square in the upper left hand corner of your YouTube video:

Because the GPU acceleration only works on NVIDIA hardware, owners of the new 15/17-inch MacBook Pros will tradeoff lower battery life for lower CPU utilization (the NV GPU has to be powered up during Flash video playback). Hopefully this is just the first step as there's no reason why Intel's HD graphics can't offer the same H.264 acceleration as the NVIDIA GPUs.

And to set the record straight, I wasn't laying on the couch with my iPad.

Now that the week of the 12th is upon us, let's see how a quick stock check at the usual suspects turns out.

Although not all of the GTX 400 series cards from the initial wave have reached etailers yet, it looks like the bulk of the cards have come in and gone out the door. Newegg, ZipZoomFly, and MWave all have had cards going in and out of stock all day long, while Best Buy and Amazon are still backordered/pre-ordering. Meanwhile pricing is highly variable with MSRP cards selling out quickly while you can still find and buy a card if you're willing to pay a $30-$50 premium.

So who has what cards and where? Read on to find out.

Today we have something a bit different for you from the GPU world: aftermarket cooling. Thermalright sent over a complete aftermarket GPU cooling kit for our Radeon 5870 containing 3 of their products: their Spitfire GPU heatsink, their VRM-R5 VRM heatsink, and their X-Silent 140mm fan, which combine to form one truly monstrous GPU cooler.  Thermalright claims that they got a 32C reduction in the load temperature of a Radeon 5870 in their labs, so let’s put this to the test and see if their latest cooler is as good as they claim it is.

Last September when AMD unveiled its new DirectX 11 GPUs, we were promised the ability to drive a minimum of 3 or a maximum of 6 displays off of a single card.  Every single AMD DX11 graphics card that has launched since has delivered on that promise, even down to ridiculously low price points.  The one thing we've been missing is a card that we can hook up to six displays.  Until now that is.

This is the Radeon HD 5870 Eyefinity 6 Edition.  It's a standard 5870 with 2GB of GDDR5 memory instead of the default 1GB, all clock speeds are the same (850MHz core, 1.2GHz memory).  Power requirements go up slightly due to the increase in memory size, and thus the card requires a 6-pin and an 8-pin power connector.  The big difference is the Eyefinity 6 Edition's 6 mini Display Port connectors which allowed us to drive one ridiculous display in today's review.

The 6 month wait for Fermi is finally over. Today NVIDIA is launching their first two Fermi family products: The GeForce GTX 480 and GeForce GTX 470. Based on the GF100 GPU, the GTX 400 series is targeted directly at the high of the market, competing against AMD’s Radeon 5800 series.

NVIDIA has promissed a great deal about the Fermi family. It's supposed to be a compute monster, a tessellation powerhouse, and of course great at games. A 3 billion transistors and over 500mm2 in die size it's certainly some kind of giant.

So how does it stack up? We've run NVIDIA's latest through our new GPU test suite to see if NVIDIA can deliver on their promises, justify their pricing, and take back their single-GPU performance crown.

To close the 5850/5770 gap, today AMD is introducing the 5830, a further cut-down Cypress. Today we'll see why this card embodies the idea that there is no such thing as a bad card, only a bad price

With the supply of AMD's Cypress chips picking up, we finally have our first fully-custom board. Today Sapphire is releasing the first such product, the overclocked and overcooled 5850 Toxic Edition.

Today we're looking at MSI's N210-MD512H, based on the last 40nm GeForce 200-series part we have yet to see: the G210.

Some big changes are coming for AMD's Catalyst drivers. We have everything you need to know about what's new today with 10.2, along with what's coming next month with 10.3.

After the RV770 ATI surely must have had an easy time building the RV870, right? Wrong. While we now know that the RV770's strategy paid off, the RV870 was being designed before the 770 ever made it out the door and building another RV770 was a tough sell internally. This is the backstory of the GPU that eventually became the Radeon HD 5870.

Since we published our reviews of the Radeon HD 5450 and the Radeon HD 5570, we have been going back and forth with AMD over the results of our video quality tests using the Cheese Slices test. Our initial results showed that neither the 5570 nor the 5450 had enough compute power to handle the full suite of post-processing abilities on 1080i video, the most important of which was Vector Adaptive deinterlacing. Since then, AMD has let us in on a few things that have changed that significantly, so let’s dive right in.

With the bulk of the 5000 series launched, AMD is now launching products to fill in the gaps in their product line. Today we look at one of those cards, the Redwood based 5570, and why it's one of fastest low profile cards on the market

Cedar, AMD's final Evergreen chip, launches today with the Radeon HD 4500. What does bringing Evergreen, audio bitstreaming, and DX11 do for the low-end of the market? We find out

At long last, the other shoe drops. Over CES NVIDIA told us everything we wanted to know about the GPU that was Fermi. We find out why NVIDIA believes it will be every bit the graphics monster that it is the compute monster, and what it will take for NVIDIA to succeed.

Today AMD is launching the first Redwood card - the 5670 - bringing DX11 down to its lowest price point yet. While the performance won't set the world on fire, does the card have enough zip where it matters?

Today the first motherboard with Lucid's Hydra technology launches: The MSI Big Bang Fuzion. Will the Hydra shake up the multi-GPU world as Lucid has been claiming?