Due to local system maintenance on Tuesday, September 27th, nanoHUB will be unable to launch simulation jobs on clusters conte, rice, carter, and hansen. We apologize for any inconvenience.
Find information on common issues.
Ask questions and find answers from other users.
Suggest a new site feature or improvement.
Check on status of your tickets.
By Gerhard Klimeck
Electrical and Computer Engineering, Purdue University, West Lafayette, IN
Write a review
Frederico Marcolino Quintao Severgnini
23 Mar 2016
5.0 out of 5 stars
As a graduate student in ECE, I find this topic fascinating. Understanding where the Moore's Law is heading to and what are the next challenges engineers will face is an exciting topic, that was presented in a simple and clear way.
Please login to vote.
Jeong Hun Kim
24 Apr 2013
I am not a semiconductor or computer architecture major so I was wondering why at some point in the past 5 years or so, the speed of a CPU was maxed out at around 3GHz and not increasing much past that. Before then, it seemed that the speed doubled every 6 months or so, so back then I was hesitant to buy a high-end CPU since I knew that there will be a faster one very soon. It was interesting to know that the main issue ( I was speculating so but wasn’t 100% sure) that the clock speed increase stopped because of power limitation, in particular switching circuits on and off in a CMOS inverter leads to power consumption. Basically, the power consumption is proportional to the frequency of the switch ( the clock speed ) and the square of the gate voltage. So, one way to reduce power consumption is to reduce gate voltage but this in turn leads to power consumption through leakage, which is proportional to the exponential inverse of the gate voltage. It was also interesting to know that now, the transistors are so small that assuming continuum is no longer valid, everything should be modeled as a finite type problem. The idea of being able to count about 40 atoms in a 5nm gate transistor was kind of astonishing to me. I think there are transistors out there being manufactured in the scale of 20~30nm currently so there are about 200 or so atoms in a transistors already. And the idea of a single phosphorus atom transistor leads to unexpected things like the Coloumb diamond and Coloumb stair cases where increasing the voltage will actually ‘decrease’ the amount of current flow since we are only dealing with a single atom. ( Which kind of reminded me that one of the articles that was posted along with the presentation was titled “Ohm’s law survives to atomic scales” or something along those lines. Doesn’t this Coloumb stair case phenomenon mean that Ohm’s law doesn’t hold? ) Also finally, I was getting the impression that the single-atom transistors were being built and did get confirmation from your presentation that through the use of tunneling microscopes, that a single atom transistor was indeed built and the Coloumb diamond phenomenon had been verified. I didn’t think that they would be able to be mass produce single atom transistors and you had also verified that during the presentation. ( At least with current technology. ) Overall, the presentation was very entertaining and informative, giving out practical knowledge that is used in real life. Even for a non-major in the field, I was able to understand the key elements and I believe that will be true for most people. I got lost on some of the middle slides since they were not in my field of research but still, I believe that I got most of the general idea that you were trying to pass around.