[A83] Re: Faster Multiplication
[Prev][Next][Index][Thread]
[A83] Re: Faster Multiplication
In accordance with the prophecy, Thomas J. Hruska uttered:
> At 04:33 PM 5/30/01 EDT, you wrote:
> >never even heard or seen anything like this before....shesh....
>
> When you get into programming groups like these, you see stuff like this.
> The reason is that the programmers here count clock cycles every time that
> they write code.
Hmm, no I don't. =) I do know that an O(log n) multiplication routine is better
than an O(n) one, but I don't count clock cycles. I'm sure some programmers do,
though.
> You've played some of the assembler games for the 83, I'm
> sure and have noticed that they are fast. Well, they are fast because the
> people here are not sloppy coders. Today's Windows games seem about as
> fast as the games on your 83...that's because of something known as bloat.
> The programmers here scrape for every byte they can save and every clock
> cycle they can save. Thus, you even have 3D games on that 6 mHz processor
> in your 83!!! (They don't run terribly fast (7-9 fps?), but it does
> demonstrate that today's 3D games could be smoother, faster, and better
> looking).
There's money involved in that. Intel pays major (and minor?) software companies
to make their games slow, and include unnecessary delay routines into them. This
way, people figure they need a faster processor, and so they buy one from Intel.
It is also worth noting that Intel probably know how to make faster processors
than they're currently selling. (This is just speculative, mind you.) The guy
who gave us Moore's Law (computing power doubles every 18 months) was the
founder of Intel, and naturally they try to stick to his prophecy as exactly as
possible.
Linus
--------------------------------------------- http://www.df.lth.se/~lft ---
Syntactic sugar leads to cancer of the semicolon.
-- Alan Perlis
Follow-Ups:
References: