PC manufacturers are starting to get worried that they cannot get enough of AMD’s new 40nm GPUs.

The problem is due to AMD’s foundry having trouble ramping up production of a new 40nm process, and it means that some PC makers are having to delay shipments of desktop and laptop computers.

It seems that the difficulty is Taiwan Semiconductor Manufacturing Company (TSMC). It can’t make enough of AMD’s 5800 series 40nm GPUs. The foundry is running flat out but yields apparently are still below expectation.

Matt Davis, a spokesman for AMD, told Computerworld that the design of the chips was sound. It was all just a matter of trying to get TSMC to a point where they can yield.

“They’re feeling the manufacturing crunch,” said Davis. “We’re a little bitter under yield but we’re working back into a manufacturing schedule we want for these parts. TSMC can only kick them out so fast at this point.”

We think he actually meant to say, “a little bit” and hope Chimpzilla’s not really bitter about it. AMD said that it hopes to have production up to speed by year’s end.

If it was a product that no one wanted then there would be no problem but demand for the chips is higher than expected. The company’s HD5800 series GPUs have the industry performance lead at the moment.

AMD has a little time to get manufacturing in line before PC vendors start looking for greener pastures. AMD’s rival Nvidia also uses TSMC and is also reporting trouble with 40nm parts yields.

Full story available at theinquirer.net

After a 20-month investigation, attorney general Andrew Cuomo accused Intel of striking deals in which computer makers such as Dell, Hewlett-Packard and IBM agreed to choose its microprocessors over those of competitors in return for billions of kickbacks disguised as “rebates”.

“In Intel’s market, there were no laws of supply and demand,” said Cuomo. “Instead, it was the law of ‘my way or the highway’.”

Cuomo’s action follows complaints by Intel’s leading rival, Advanced Micro Devices (AMD), over unfair tactics in sales of microprocessors. The charges are along similar lines to those laid by the European Commission, which imposed a €1.03bn (£922m) fine on Intel in May, in the biggest punishment ever enacted by Brussels for anti-competitive behaviour.

“Rather than compete fairly, Intel used bribery and coercion to maintain a stranglehold on the market,” said Cuomo. “Intel’s actions not only unfairly restricted potential competitors, but also hurt average consumers who were robbed of better products and lower prices.”

Click here to read full story.

So, why would Apple adopt a Core i processor in a laptop? The short answer is OS X Snow Leopard. The new operating system is designed to be better at wringing more performance out of multicore processors–and the Core i chips pack four cores.

The long answer is the Core “i” chips themselves. The Core i, a.k.a. “Nehalem,” is a brand new Intel microarchitecture brimming with performance improvements over the current Core 2 design.

For starters, the recently announced mobile i7-820QM processor integrates a hefty 8MB of cache memory–compared with the 6MB in the fastest Core 2 Duo that Apple currently offers on the MacBook Pro. Generally, the larger the cache memory, the better the performance.

But Intel has done a lot more than just up the transistor count via a larger cache. While the quad-core i7-820QM is rated at 1.73GHz, a single core can be “overclocked” to 3.06 GHz. Intel does this with a technology dubbed Turbo Boost, which speeds up and slows down individual cores to meet processing and power-efficiency needs, respectively.

Need more convincing? The Core i7 also comes with Hyper-Threading, which can double the number of tasks–or threads–a processor can execute. You won’t find that in the Core 2 chips.

Digital media creation also gets a boost. Intel claims up to 81 percent faster video encoding.

Click here to read the full story at cnet.com.