by Anton Shilov
10/26/2007 | 11:31 PM
Volume production of a new generation of microprocessors for desktop personal computers, laptops, servers and other computing devices officially began today inside of Intel Corp.’s first high-volume 45nm manufacturing factory in Chandler, Arizona. The first 45nm chips, which were produced on Intel’s D1D fab in Hilsboro, Oregon, will be unveiled next month and will serve high-performance desktops and mainstream servers.
“The opening of Fab 32 in Arizona today is a testament to Intel’s continued investment in our most strategic asset – the most advanced, environmentally friendly manufacturing network in the world. The magic of 45nm and our new transistor design allow us to deliver high-performance, energy-efficient processors to our customers across the entire spectrum of market segments, from the most powerful servers to a variety of mobile devices and everything in between,” said Paul Otellini, Intel president and chief executive.
Fab 32 is Intel's sixth 300mm wafer factory and its second factory, after D1D in Oregon, to produce 45nm chips. Two additional 45nm, 300mm manufacturing factories are scheduled to open next year in Kiryat Gat, Israel (Fab 28) and Rio Rancho, New Mexico (Fab 11x). Using 300mm wafers lowers the production cost per chip while diminishing overall use of resources.
Intel’s Fab 32 structure measures 1 million square feet and has 184 thousand square feet of clean room space. More than 1000 employees will operate the factory in such positions as process, automation and yield engineers and senior manufacturing technicians.
Intel is the first chipmaker to implement an innovative combination of new materials that reduces transistor leakage and increases performance in its 45nm process technology. The company will use a new material with a property called high-k, for the transistor gate dielectric, and a new combination of metal materials for the transistor gate electrode.
The first processors to utilize the new process technology are code-named Penryn, Wolfdale and Yorkfield. The first one is aimed at mobile computers and will be officially unveiled early next year, whereas Wolfdale and Yorkfield are dual-core and quad-core processors aimed at desktops and servers. Just like today, Yorkfield chips will use two dual-core Wolfdale/Penryn dice. Typically, the whole family of 45nm microprocessors from Intel is called Penryn.
Intel announced in March that the new lineup of chips produced using 45nm process technology will have greater instructions per clock (IPC) execution, which means that they will be faster and more efficient even at the same clock-speeds with the current generation chips. Besides, the new chips will be able to run at higher clock-speeds compared to today’s Core 2 Duo and Core 2 Quad products.
The major micro-architectural improvements for new Intel Core 2 processors, besides SSE4 instruction set, include the so-called Unique Super Shuffle Engine and Radix 16 technique. The Super Shuffle Engine is a full-width, single-pass shuffle unit that is 128-bits wide, which can perform full-width shuffles in a single cycle. This significantly improves performance for SSE2, SSE3 and SSE4 instructions that have shuffle-like operations such as pack, unpack and wider packed shifts. This feature will increase performance for content creation, imaging, video and high-performance computing. Radix 16 technique, according to Intel, roughly doubles the divider speed over previous generations for computations used in nearly all applications. In addition, Intel also improved virtualization technology as well as added some features to dynamic acceleration technology, which is supposed to boost single-threaded applications’ performance on multi-core chips.
Each of Intel’s dual-core Penryn chips will have 410 million transistors, up significantly from 291 million of current dual-core Conroe processors, however, thanks to 45nm process technology, the chips will have die size of 107 square millimeters, down about 25% from 155 square millimeters of the Conroe.