Previous | Next --- Slide 27 of 43
Back to Lecture Thumbnails

The reason why we could simply increase clock frequency to obtain performance is because the size of transistors continued to decrease during that period. Smaller transistors have a smaller delay between states than larger ones. If power and heat wasn't a concern, I bet we would just continue increasing clock frequency until transistors stop getting smaller.


@leaf2fire Then we probably reached a philosophical question, how small can a transistor get?

Just for fun, what if quantum computers becomes practical for daily usage? While transistor technology may not be rapidly evolving, quantum computing could possibly provide another form performance buff that far exceeds performance gains from normal transistor based chips. I mean how fast do we need our computers to be?

Quantum computing a reality?


@Funky9000, I think the conventional view is that quantum computers will have a small range of applications where they will make huge changes to the state of the art, but most applications will not be affected. If your daily usage involves factoring large numbers or doing optimization problems, this might be useful. Our current end-user applications don't seem to need these capabilities (though it's entirely possible that we'll find them very useful if they become cheap and available).


@Funky9000 We are most likely limited by the size of "atomic stuff." Clearly I'm not an electrical engineer - how far are we from this? In a brief search I found that the smallest we have done so far is roughly 4nm (which consisted of 7 atoms together).

How much smaller do people predict transistors getting?