AMD Explains Advantages of High Density (Thin) Libraries

We have seen the improvements that the "Steamroller" will offer in performance per watt with its design improvements. In addition to those improvements, AMD will be using "dense" or "thin" libraries employed by its GPU design teams, but for CPU implementation.

AMD told us that products currently shipping with 32nm use a combination of automated place and route and hand-placed semi-custom design (top plot), which reduces power and area somewhat. To deliver more power efficient computations, AMD has employed a high-density cell library to reduce the area and power by 30 percent (bottom plot). The design yields a more portable and energy efficiency CPU core employing industry standard design methodologies well adapted to a foundry model. These improvements, according to AMD, are yielding a 15 to 30 percent lower energy per operation for power constrained designs, as compared to a full process node improvement.

Look for more details from AMD during Hot Chips Symposium on its Surround Computing and Steamroller.


Contact Us for News Tips, Corrections and Feedback

This thread is closed for comments
    Your comment
  • wiyosaya
    In some respects, this sounds like the programmable gate array concept. It is interesting to see this adapted to non-programmable chip design.

    I am somewhat surprised, though, that this implies that such optimization was never before computerized. I would be really surprised if there were no computer optimization of chip layouts before this.

    So, is this just AMDs marketing engine at the helm again?
  • madooo12
    read on anandtech that they will only be used in excavator
  • Ragnar-Kon
    So more logic in a smaller area. Basically what chip designers have been doing since ICs were first invented. Nothing new...

    EDIT: My bad, more logic in a smaller area without a die shrink. So essentially just housecleaning on current libraries. Still clever marketing.
    It is a die shrink, myyyy baaddddd. Doesn't seem anything like the 3-D transistors used in Intel's 22nm process though. Not that is necessarily a bad thing, I just thought it was similar to that originally.