Selected Nvidia partners will unveil Nvidia's next-generation GF100 based flagship, the Geforce GTX 480 at Cebit, which runs from March 2nd to March 6th. Official release is expected towards the end of March, though widespread availability might slip to April, assuming there are no more mishaps.

More details next page.

The new ASUS laptops are going to be equipped with NVIDIA Optimus Technology that allows seamless transition between discrete and integrated graphics for enhanced viewing and longer battery life. The notebooks are scheduled to be available from March.

Update: Nvidia have now announced on their Twitter page that GF100 will be branded Geforce GTX 480 and Geforce GTX 470.

Long rumoured to be codenamed GT300, with branding as the GTX 300 series, Expreview is now reporting that Nvidia has named GF100 as Geforce GTX 480 for the top model and GTX 470 for the more affordable, cut down variant.

Release is on track for late Q1 2010, i.e. end March, barring further delays. Why Nvidia would seemingly skip a generation is unknown, though an upgrade from GTX 285 to GTX 480 does sound psychologically more significant than to GTX 380. Considering GF100 is a revolutionary architecture for Nvidia, we could even expect a completely new branding scheme, like they did for the GT200 series.

More next page.

TSMC's Senior VP of Operations, Mark Liu, has mentioned that TSMC's troubled 40nm process is now of the same quality as the mature 65nm process. The chamber mismatch issue which dropped the yields of the HD 5800 series to a reported 40% have now been resolved, and yields are recently rumoured to be between 60% and 80%.

Mr. Liu did not delve into further details. Yield rates are a win-win-win situation for everyone involved. ATI will be able to ship more GPUs at lower manufacturing costs, customers will get widespread availability and lower prices.

More next page.

Fudzilla has reported the existence of a dual-GPU version of Nvidia's upcoming GF100 in April, roughly one month after the single-GPU GF100 launch, if all goes well and A3 is the final stepping. However, is such a mammoth product, the dual-GF100, actually feasible?

Earlier rumours pointed to a TDP of 180-200W for the single GPU GF100, but the recent demonstration in CES 2010 saw a 8 pin + 6 pin PCI-e power configuration, suggesting a TDP of somewhere between 225W and 300W. Since then, SemiAccurate have reported the actual TDP to be a rumoured 280W.

More next page.

Much like the Radeon HD 4890 last generation, ATI is rumoured to be preparing a refresh for its Cypress GPU. Details are unavailable at this stage. Whether it will be a die shrink, increase in shader counts, or just clock speed optimizations (ala HD 4890) is unknown.

The HD 5800 refresh should come in right after the GF100 as a direct competition. If GF100 does end up competitively priced, ATI have a significant price cut buffer in the short term, with the refresh taking over within a couple of months.

"7" might not be the most oft-used number for naming products this side of the globe, but Gigabyte has chosen to use the "UD7" moniker on it's upcoming feature-packed mainboard. Today we landed our hands on a pre-production sample, so here we are with the photos.

There has been considerable speculation of yields for TSMC's troubled 40nm process. From single-digit percentage yield rumours of early steppings of Fermi to under 40% for Cypress, there has been very little positive reports from TSMC's process. TSMC themselves admitted yields hovering around 40% in November, due to "Chamber mismatch issues". They promised to fix these issues by the new year, which they have sufficiently delivered on - as evidenced by the sudden widespread availability of the HD 5800 series, which was previously negligibly available.

This is reflected by the latest yield figures as suggested by Fudzilla's sources. Up from 40% to at least 60% and as much as 80%.

More next page.

In a curious report from TomsHardware, an Nvidia partner claims Nvidia's next generation Fermi chips will go into production in the 3rd week of February, with limited quantities available in mid-March.

It is not clear if this is the desktop (GF100) or the Tesla version of Fermi. The source goes onto say that 40nm TSMC yields are expected to be poor, though better than ATI's Radeon HD 5800 series.

More next page.

Fresh from Nvidia's premiere of a GF100 working prototype, the dual-GPU variant of GF100 is now rumoured to be released some time in May 2010, at earliest. Not surprisingly, the dual-GPU, if it releases, will be a good one or two months after the single-GPU GF100 variants.

However, the feasibility of such a product is very much debatable. While originally the GF100 was expected to have a TDP of ~200W and a 6+6 pin power configuration; the prototype demonstrated features 8+6 pin - the maximum official PCI configuration. It is likely that the final stepping will be significantly more energy efficient, however.

More next page.