DARPA circuit achieves speeds of 1 trillion cycles per second

phys.org

30 pointsantoinek5 comments

That's certainly an impressive frequency, but it's far easier (relatively speaking) to construct an analogue amplifier for high frequencies than it is to, e.g., make a logic gate or counter that works at those frequencies - which would be necessary to build something like a CPU. So it's likely that even when/if this technology shows up in mass-produced analogue circuits, digital computers sequentially executing 1 tera-instructions-per-second are still a long way off.

And...

gain above 1 THz (1012 GHz)

one terahertz (1012 GHz)

lolwut? Are they trying to define 1THz = 1012GHz?

Like you, in the "caption"y thing at the top, I see

gain above 1 THz (1012 GHz)

However, in the first paragraph, I see

one terahertz (10^12 GHz)

Of course, one terahertz should be 10^12 Hz. So perhaps the author of the phys.org article got very confused.

Or just rearranged his sentence at some point and forgot to change units.

above 1THz.

Think of it this way: the speed they actually tested at was 1012 GHz. This is a speed above 1THz.

Pretty sure 10<sup>12</sup> indicates that they're talking 10 to the 12th power. Seems like they just needed one more proofreader.

Reply on news.ycombinator.com