Intel Future Processor Architecture

Intel's Future Processor Architectures

Insanely powerful chips in unbelievably small packages is what this new family of processors based on the 32 nm fabrication process are supposed to offer. Let's see what makes them tick.

After launching the meanest desktop class processor, Intel's next big step is to shrink the chip. Currently in production and available in stores, the Core i7 family of processors is the latest from Intel and sports a new microarchitecture. Nehalem as it is known to the world, comes with a number of achievements; it is Intel's first native quad core processor, has an inclusive level 3 cache, an integrated memory controller and the QuickPath Interconnect. But what was still similar to its predecessor, the Penryn architecture, was the 45nm process.

Direct Hit!

Applies To: Everyone USP: Learn about the future microarchitectures from Intel Primary Link: www.intel.com/ pressroom/kits/32nm/westmere/index.htm Google Keywords: Intel Westmere

Now the 45nm semiconductor manufacturing process isn't really a drawback for Nehalem. Intel follows a tick tock model for its products. Every tick is a shrink of the manufacturing process and tock is a brand new microarchitecture. Therefore, the 45 nm process was an intentional component for Intel while launching Nehalem as it helps them follow the model and come out with new products more quickly. Now the Westmere microarchitecture, as Intel calls it, will be the successor to Nehalem.

Westmere will be the 32nm shrink of the Nehalem microarchitecture and will have some new features. Are you wondering what will happen when the most powerful processor on this planet migrates to a 32nm process? Well, let's have a look.

Increased power efficiencyWhen you build Nehalem on a smaller manufacturing process and throw in a new feature or two, you get Westmere. A smaller manufacturing process means that Westmere is going to be more power efficient and cost effective as with a shrink in size, Intel can manufacture more cores per wafer.

AES instruction setSimilar to the 45nm tick of Penryn, Intel plans to add new capabilities to 32nm processors using the additional transistor budget. The 32nm Westmere processors offer new Advanced Encryption Standard (AES) instructions. The new instruction set consists of 7 instructions for accelerating encryption and decryption. Early Intel analysis shows a significant performance boost on encryption/decryption algorithms. An example of client usage for this can be full disk encryption.

A new multi-chip packageBoth, a processor core and an integrated graphics and memory controller on a single processor will be seen with the first Intel 32nm Westmere products. Interestingly, even though the processor core will be 32nm, the integrated graphics and memory controller will be on a separate 45nm die. Since the integrated graphics and memory controller are located on a separate die, rumors are you can adjust power settings and clock frequencies better than on one die.

Repartitioned mainstream client platformWith Westmere, the client side platform gets repartitioned into a 2 chip solution as compared to a 3 chip solution present in Penryn. The new 32nm Westmere-based platform features a two chip solution with just the processor and an Intel 5 series chipset. As a result, Intel has eliminated the need for an ICH, which reduces manufacturing costs and valuable space on the motherboard. Also, with higher degree of integration there is a natural performance enhancement while lowering power consumption.

Coming soon but not 32nmCore i7 might be advantageous to professionals running a large number of multi-threaded applications, but for routine app users, such amazing performance is hardly needed. Moreover gamers need higher clock speeds rather than too many cores and thread support from their processors.
Slated to hit markets during Q3 this year, Intel has two new 45 nm processors based on the Nehalem microarchitecture. These are Lynnfield (desktop) and Clarksfield (laptops). Likely to be launched as Core i5, Lynnfield will be be the first mainstream Nehalem, following on the successful launch of the Core i7 line for enthusiasts. Like Core i7, Core i5 will be based on 45nm process, will have 4 cores and hyper-threading, so each core can handle two processing threads.

The 32nm processorsAfter the launch of 45nm Lynnfield and Clarksfield processors, we should see the first 32nm processors. The first generation of 32nm products from Intel shall be aimed at desktop and mobile users. Clarkdale for mainstream desktop and Arandale for laptops shall be the names of these processors.

They will have two cores supporting up to 4 threads and will also have integrated graphics. Interestingly, Intel doesn't have a 32nm quad core processor on their roadmap as of now. Desktop users will see the launch of Piketon and Kings Creek platforms utilizing a new Intel 5 series chipset and socket with Lynnfield or Clarkdale processors. Both desktop processors will need the new Intel socket LGA 1156 and one of the upcoming 5-series chipsets (P55, P57) in order to run. The X58 Express platform that was launched with Nehalem will also see an update later this year. This will be a 32nm 6-core processor supporting 12 threads and is codenamed Gulftown.

When talking of 32nm processors, how could server processors be left behind. Intel is planning to introduce 32nm Xeon processors for the entry level clients and then ramping them up to the high-end clients with time. At entry level, which usually consists of single socket servers, a platform called Foxhollow will bring the Lynnfield processor (designed based on Nehalem) into this arena, to be followed later by the 32nm Clarkdale processor. For dual-socket servers, the Nehalem Efficient Performance (EP) processor is expected in Q1 2009, this will be refreshed next year with a 32nm Westmere-based product in future, as will Nehalem EX for four-socket systems. The expandable server market should see Nehalem-EX followed by 32nm Westmere-based processors next year.

ConclusionIntel's tick tock model has delivered new technology and processor micro architectures, regularly. Westmere is just a change in the process technology whereas the next major change will come with a 'tock', the Sandy Bridge. What Westmere also means for mainstream consumers is that the long lived socket LGA 775 is seeing its replacement for the very first time. It is hard to believe that LGA 775 platforms have been around since 2004 and were able to last to this point. But they will soon be replaced with the new socket LGA 1156, along with the upcoming 5-series of chipsets
(P55, P57).

Mobile phone brain cancer link rejected

Further research has been published suggesting there is no link between mobile phones and brain cancer.

The risk mobiles present has been much debated over the past 20 years as use of the phones has soared.

The latest study led by the Institute of Cancer Epidemiology in Denmark looked at more than 350,000 people with mobile phones over an 18-year period.

Researchers concluded users were at no greater risk than anyone else of developing brain cancer.

The findings, published on the British Medical Journal website, come after a series of studies have come to similar conclusions.

'Reassuring'

But there has also been some research casting doubt on mobile phone safety, prompting the World Health Organization to warn that they could still be carcinogenic.

In doing so, the WHO put mobile phones in the same category as coffee, meaning a link could not be ruled out but could not be proved either.

The Department of Health continue to advise that anyone under the age of 16 should use mobile phones only for essential purposes and keep all calls short.

Continue reading the main story

“Start Quote

These results are the strongest evidence yet that using a mobile phone does not seem to increase the risk of cancers of the brain or central nervous system in adults”

Hazel NunnCancer Research UK

The Danish study, which built on previous research that has already been published by carrying out a longer follow-up, found there was no significant difference in rates of brain or central nervous system cancers among those who had mobiles and those that did not.

Of the 358,403 mobile phone owners looked at, 356 gliomas (a type of brain cancer) and 846 cancers of the central nervous system were seen - both in line with incidence rates among those who did not own a mobile.

Even among those who had had mobiles the longest - 13 years or more - the risk was no higher, the researchers concluded.

But they still said mobile phone use warranted continued follow up to ensure cancers were not developing over the longer term, and to see what the effect was in children.

Hazel Nunn, head of evidence and health information at Cancer Research UK, said: "These results are the strongest evidence yet that using a mobile phone does not seem to increase the risk of cancers of the brain or central nervous system in adults."

Prof Anders Ahlbom, from Sweden's Karolinska Institute, praised the way the study was conducted, adding the findings were "reassuring".

Prof David Spiegelhalter, an expert specialising in the understanding of risk who is based at the University of Cambridge, said: "The mobile phone records only go up to 1995 and so the comparison is mainly between early and late adopters, but the lack of any effect on brain tumours is still very important evidence."

And Prof Malcolm Sperrin, director of medical physics at Royal Berkshire Hospital, said: "The findings clearly reveal that there is no additional overall risk of developing a cancer in the brain although there does seem to be some minor, and not statistically significant, variations in the type of cancer."

But the researchers themselves do accept there were some limitations to the study, including the exclusion of "corporate subscriptions", thereby excluding people who used their phones for business purposes, who could be among the heaviest users.

http://www.bbc.co.uk/news/health-15387297

DNA COMPUTING

DNA computing is a form of computing which uses DNA and molecular biology, instead of the traditional silicon-based computer technologies.

This field was initially developed by Leonard Adleman of the University of Southern California. In 1994, Adleman demonstrated a proof-of-concept use of DNA as form of computation which was used to solve the seven-point Hamiltonian path problem. Since the initial Adleman experiments, advances have been made, and various Turing machines have been proven to be constructable.

There are works over one dimensional lengths, bidimensional tiles, and even three dimensional DNA graphs processing.

On April 28, 2004, Ehud Shapiro and researchers at the Weizmann Institute announced in the journal Nature that they had constructed a DNA computer. This was coupled with an input and output module and is capable of diagnosing cancerous activity within a cell, and then releasing an anti-cancer drug upon diagnosis.

DNA computing is fundamentally similar to parallel computing in that it takes advantage of the many different molecules of DNA to try many different possibilities at once.

For certain specialized problems, DNA computers are faster and smaller than any other computer built so far. But DNA computing does not provide any new capabilities from the standpoint of computational complexity theory, the study of which computational problems are difficult. For example, problems which grow exponentially with the size of the problem (EXPSPACE problems) on von Neumann machines still grow exponentially with the size of the problem on DNA machines. For very large EXPSPACE problems, the amount of DNA required is too large to be practical. (Quantum computing, on the other hand, does provide some interesting new capabilities).