Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
14 years ago

40nm vs 60nm Power

Hello,

I had been under the assumption that a 40nm device will consume less power than a 60nm device given the same code. I tried that with an Arria II GX (40nm) and a Cyclone IV GX (60nm) and I found that give the same design and test parameters, the Arria II GX will consume more power than the Cyclone IV GX. I got this using the PowerPlay power analyzer.

Can anyone shed a little light on this?

Thanks!

2 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    Reduction of power with process shrinkage is a general trend, but it's neither automatic or guaranteed.

    Leakage (static power) can sometimes increase.

    Dynamic power reduction usually comes via voltage reduction, with is not the case: both Cyclone IV and Arria II are 1.2V VCCINT.

    Additionally, the Cyclone family design puts a greater emphasis on power than Arria.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    Thanks for the answer rbugalho!

    Arria II is actually 0.9V. I expected that to help the power consumption. Below is the summary for both. As expected, the dynamic power is greater on the Cy IV. I was surprised though that Arria II uses a lot more power for I/O.

    Arria II:

    Total Thermal Power Dissipation: 2346

    Core Dynamic Thermal Power Dissipation: 770

    Core Static Thermal Power Dissipation: 887

    I/O Thermal Power Dissipation: 688

    Cy IV:

    Total Thermal Power Dissipation: 1894

    Core Dynamic Thermal Power Dissipation: 1263

    Core Static Thermal Power Dissipation: 448

    I/O Thermal Power Dissipation: 183