Moore's Law Marches on at Intel

Intel President and CEO Paul Otellini today displayed a silicon wafer containing the world's first working chips built on 22nm process technology. The 22nm test circuits include both SRAM memory as well as logic circuits to be used in future Intel microprocessors.

"At Intel, Moore's Law is alive and thriving," said Otellini. "We've begun production of the world's first 32nm microprocessor, which is also the first high-performance processor to integrate graphics with the CPU. At the same time, we're already moving ahead with development of our 22nm manufacturing technology and have built working chips that will pave the way for production of still more powerful and more capable processors."

The 22nm wafer displayed by Otellini is made up of individual die containing 364 million bits of SRAM memory and has more than 2.9 billion transistors packed into an area the size of a fingernail. The chips contain the smallest SRAM cell used in working circuits ever reported at .092 square microns. The devices rely on a third-generation high-k metal gate transistor technology for improved performance and lower leakage power.

By continuing to lead in manufacturing technology Intel is able to innovate and integrate new features and functions into its processors. Intel's 32nm process is now certified and Westmere processor wafers are moving through the factory in support of planned fourth quarter revenue production. Following the move to 32nm Intel will subsequently introduce Sandy Bridge, Intel's next new microarchitecture. Sandy Bridge will feature a sixth generation graphics core on the same die as the processor core and includes AVX instructions for floating point, media, and processor intensive software.

Source: Intel (news : web)

   

• Husky - Sep 22, 2009
  • Rank: 1 / 5 (2)
  What the ultimately should do is make a really large L3 cache on die (like 1 GB) and share this with 2 cpu core and 2 ondie gpu running in crossfire, now that would blow anything out there out of the water, you could run your basic OS AND 3D physics inside the L3 and the onboard cards(wich up until now tend to suck, for stealing your slower DRAM instead of using its own VRAM)could have memoryspeeds comparable/exceeding the latest VRAM versions but at that communicate much faster with the cpu cores than an offboard card(s) talking over PCIe
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• nxtr - Sep 22, 2009
  • Rank: 4 / 5 (4)
  Funny how the nay sayers always seem to think that innovation has run its course.
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• diva4d - Sep 23, 2009
  • Rank: 1 / 5 (1)
  I wonder how this article will read in 15years' time...
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• RayCherry - Sep 23, 2009
  • Rank: not rated yet
  diva4d:
  

  I wonder how this article will read in 15years' time...

  
  Try five years. Do you remember the Pentium MMX (MultiMedia eXtensions) chips from a decade ago? You never hear MMX mentioned anymore, but those extensions are still in each chip, and evolve in parallel with DirectX in the Operating Systems.
  
  Now that GPU technology is stablizing, it too can be partially, or completely, integrated into the same package.
  
  Just hope that Intel does not return to the MMX 'slot/cartridge' form factor.
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• superhuman - Sep 23, 2009
  • Rank: not rated yet

  Funny how the nay sayers always seem to think that innovation has run its course.

  
  
  The death of Moore's law is certain. They are working on a 22nm process now and the silicon atom is 0.1nm across, there may be a few more generations and that's it.
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• rincewind - Sep 23, 2009
  • Rank: not rated yet

  Funny how the nay sayers always seem to think that innovation has run its course.

  
  
  The death of Moore's law is certain. They are working on a 22nm process now and the silicon atom is 0.1nm across, there may be a few more generations and that's it.

  
  
  Intel is confident they can sustain Moore's law until 2022.
  
  The question is whether the next generation of transistor technology (e.g. 3d molecular computing) can continue the exponential trend.
  
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• thales - Sep 25, 2009
  • Rank: not rated yet

  The death of Moore's law is certain. They are working on a 22nm process now and the silicon atom is 0.1nm across, there may be a few more generations and that's it.

  
  
  True, but - to state the obvious - exponential advances in computing power are happily not limited to integrated circuit technology. The brain demonstrates that there are other approaches to ridiculous amounts of processing power.
  
  Quantum computing is likely many decades away. I think as we approach the limits of integrated circuits we will be forced to find increasingly efficient ways of "parallelizing" (for lack of a better term) software code in order to maintain the current rate of advance. Intel is a hardware company - I wouldn't be surprised if they're already working hard on finding ways to use hardware to do parallel processing on linear code.
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• Ulg - Sep 25, 2009
  • Rank: not rated yet
  While there are limitation to size in respect to atoms- that is only half the story. The materials we use are also incredibly important. Even if you used say diamond instead of silicon wafers you would have what, 250 times the amount heat transference- that means you can bump up the clock speed significantly. If anything the real question is- when it fails will it be because it did not meet expectation or because it shattered it.
  
  remember its not just the size- its how you use it ;P
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• Scryer - Sep 27, 2009
  • Rank: not rated yet
  Moore's law has an end - the final spectrum of technology, there happens to be such a thing as knowing how to compute the fastest. We're just not there yet.
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5
• guiding_light - Oct 14, 2009
  • Rank: not rated yet
  I remember reading Intel will spend 7 billion to upgrade fabs to 32 nm in 2010. I think this is the reason they delayed immersion lithography for 45 nm. So the 32 nm transition will be quite painful, initially. But after 2010, it should be fine. Other companies seem to have weathered it ok.
  • report abuse
  • • Current rank
    • 1
    • 2
    • 3
    • 4
    • 5