Proof that 'slacking' is more productive in terms of computational power, ie Moore's Law in practice.

[Analog]

Abstract:
We show that, in the context of Moore's Law, overall productivity can be increased for large enough computations by `slacking' or waiting for some period of time before purchasing a computer and beginning the calculation.

According to Moore's Law, the computational power available at a particular price doubles every 18 months. Therefore it is conceivable that for sufficiently large numerical calculations and fixed budgets, computing power will improve quickly enough that the calculation will finish faster if we wait until the available computing power is sufficiently better and start the calculation then.

<see graphs>

link

 

[Marshallj]

linky no worky

 

[WinkOsmosis]

Originally posted by: Marshallj
linky no worky

It works fine. And that is very fascinating.

 

[RyanM]

Originally posted by: WinkOsmosis

Originally posted by: Marshallj
linky no worky

It works fine. And that is very fascinating.

That's quite fasacinating.

I'd like to see the following figured out:

Take a similar situation, with a set computer budget and a large calculation, but instead of waiting X time to buy the entire group of computers...

Buy 1 computer in the first Y months, expending x% of the budget, use those to calculate...
Buy another computer the next Y months, expending the same amount as last time, and add those computers to the previous ones...
Repeat until the money in the budget is gone.

Instead of wasting all of that time, you're getting SOME work done, and constantly adding to the computing power over time.

 

[XZeroII]

Kinda off topic: I thought moore's law delt with the number of transistors, not computational power. I know the 2 seem to more together (more transistors usually means more speed), but that's not always the case.

 

[RyanM]

Originally posted by: XZeroII
Kinda off topic: I thought moore's law delt with the number of transistors, not computational power. I know the 2 seem to more together (more transistors usually means more speed), but that's not always the case.

There's an ongoing debate about this. As far as I'm concerned, when I say Moore's law, I am saying "the number of MIPS in a CPU of X dollars, in 18 months, will have roughly doubled."