Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Cool analogy. Makes a great reference point.

As an ASIC guy, I like to occasionally casually mention to software guys that at 3GHz, light travels about four inches in one clock cycle, and it frequently really blows their minds.



Worked on me. I had to do the math.

180,000mi/sec * 5280 * 12 = 11404800000 / 3,000,000,000 = 3.8inches.

Extending that a little further... on a 45mm i7 chip going from one end to the other and back would be ~3.5inches of travel. Gives me an idea of how much of a constraint packaging is.


Didn't get any of that so had to redo in metric.

speed of light = 3 x 10^8 m/s

3 GHZ = 3 x 10^9 /s

So 3 x 10^8 / 3 x 10^9 = 0.1 metre = 10 centimetre = okay, 3.9 inches

But then metric always saves your ass. Back in physics class, they'd ask you how deep the well was if you dropped a stone & heard the water splash in 10 seconds. Before the American students could even begin their work, all the Indians would yell "500 metres!" And that's cause the gravitation constant is 10, so one half 10 times 10 times 10 is 500.


Generally agreed, but for this case I remember the speed of light as a foot per nanosecond, and the arithmetic is just dividing 12 by 3.


Maybe the Americans were busy incorporating the delay due to the finite speed of sound.


Is it just me that finds this suggestion offensive. Maybe the Indians had already thought about that and since they were already taking g as 10 m/sec^2, they thought that the difference because of sound delay would be insignificant. As would be the variability in speed of sound near the water surface due to moisture.


My response was designed to highlight the equally offensive suggestion that "all the Indians" were metric geniuses, while American students were all slow and ignorant of the metric system.


Too bad the gravitation constant is 9.8


Actually the gravitation constant is 6.67×10^-11 N m^2/kg^2. Gravitational acceleration on earths surface is on average 9.81 m/s^2, and not really a constant (it varies slightly from place to place).



A Core-i7 2600 has a die size of 216mm^2.[1] Assuming it's close to a square, that's about 14mm on each side. Electrical signals in a CPU travel almost at c, so at 3.4Ghz, signals can go across the die 6 times.

It's probably much less than that because paths in the CPU aren't straight lines. Also the i7 is a 4-core die, so it's unlikely that any signals need to go across the entire die in one clock. (Besides the clock signal, of course).

[1] http://en.wikipedia.org/wiki/List_of_Intel_Core_i7_microproc...


Back when I took a VLSI class my professor showed an awesome animation of clock distribution at various frequencies. http://www.research.ibm.com/people/r/restle/Animations/DAC01...

Check out the animations under "Clock Distributions" (numbers 19 to 25. It shows how the propagation of the clock across the die is affected by frequency (When a part of the fabric/tree is up or down it represents 1 and 0 respectively).

My favorite is the SymTree and Non-Uniform SymTree (23 and 24) which shows how having a non uniform load at various parts of the chip and a fixed tree structure affects how long each part of the chip is at a 1 or 0 state (i.e. a low load part of the chip will spend more time at a stabilized 1 or 0, while a heavy capacitive load at one part can not even reach a true 1 or 0).


propagation of an electromagnetic field in copper is actually only 75% of c, it's a significant factor in processor speed


Btw, you can drop this into Google: lightsecond / 3e9 in inches: 1 lightsecond / 3e9 = 3.93428423 inches




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: