[SOLVED] [moved] How to approximate hardware energy for different circuit technology?

[mumichang]

Hi,

I have designed/synthesized a new hardware component based on 45nm circuit technogolgy and so have measured its power/energy value. I would like to compare it against other competitive ones whose energy value is released only for 90nm tech.
Since I do not have 90nm standard cell library, I would like to approximate the energy value that I collected to the one for 90nm, or the other way around.

Is there a reasonable way for this approximation? or Does anybody know any relevant literature?

Thanks.

 

[vGoodtimes]

Re: How to approximate hardware energy for different circuit technology?

This should be in the ASIC forum. In general, "nm" no longer makes sense -- you need to specify a specific tech. 45nm could be a high-perf or a low-power or a balanced tech.

 

[mumichang]

Re: How to approximate hardware energy for different circuit technology?

This should be in the ASIC forum. In general, "nm" no longer makes sense -- you need to specify a specific tech. 45nm could be a high-perf or a low-power or a balanced tech.

Thanks for your reply. I am currently using tcbn45gsbwptc IP from TSMC. Since I am a beginner, I cannot tell if it is high-perf or a low-power or a balanced tech. I hope this info is helpful.

 

[ThisIsNotSam]

Re: How to approximate hardware energy for different circuit technology?

There is no accurate way of converting from one tech to another. For a ballpark you can look at ITRS predictions. FYI, there are free 90nm libraries from synopsys, and free 15nm libraries from nangate. You can synthesize with those and get a relatively good estimate.