The variable bit cpu

base

Why switch from binary to base 4.

It doesn't matter. The data bits and data marker bits can be detected in binary just fine :D

Bye, Skybuck.

base

Not to me, I give up, what's the right answer according to you, which is ofcourse wrong anyway since binary is perfectly suited to detect the data bits and marker bits ;)

Bye, Skybuck.

Why not, if there is an advantage? OTOH, why encode in base four and throw away half of your information?

You *still* haven't a clue.

Binary works. Why are you wasting half of your information and bandwith?

In article , Skybuck Flying wrote: [...]

They now have plastic semiconductors.

data

For flexibilitttttyyyyyyyyyyyy ohhhhhhhhhhhhhhhhhh :)

just

have

the

arithmetic

Are you going to say what base you think is suited lol ? or are you just going to continue playing mister bad troll lol :D

No, I'm saying that you haven't a clue. Hint: You're wasting half your information in flags. Some information is necessarily lost in such schemes, but it need not be half. You have two bits for every bit of data. That's rather poor. An idiot could improve that by 50%.

Troll? Me? No, that's your job here.

See your doctor. Viagra might help.

In article , keith wrote: [...]

Actually, he is throwing away more than 1/2 the band width. He wastes an extra 2 bits on every positive number. This is a bummer because the way most software gets written most of the numbers are positive.

The real trouble will happen when he attempts to code the floating point operations. I'm sure he hasn't seen how sticking with this silly base 2 system gets him into trouble on that.

than

that

what

you

in

Dude what does base have to do anything with it.

To calculate the ammount of storage needed one can simply use base 2.

( Log10( value ) / Log10( base ) ) * 2 = digits needed

Though squared root 2 could be used as a base as well.

Log10( value ) / Log10( squared root(2) ) = digits needed

However I think all arithmetic will still take place in base 2.

So I do not see the relevance of base squared root 2.

Care to explain ? ;)

Bye, Skybuck.

I read in sci.electronics.design that Skybuck Flying wrote (in ) about 'The variable bit cpu', on Sat, 20 Aug 2005:

Indeed, but why introduce 10? The equation is true for any base of logs, and using base 2 logs tends to simplify it.

What is the subtle difference between 2 and (root 2) squared?

The optimum base for computing can be shown to be 'e' (2.718..) but no-one knows how to do that. But 3 is theoretically better than 2, and can be implemented in ICs as -1, 0, +1. Intelligent beings elsewhere may use this system.

My code is not about compression. If you want to compress stuff there are other much more efficient codes for positive integers only ;)

Besides from that the sign bit if that's what you mean is a question of interpretation. Since I don't like complex two's complement ideas and arithmetic my cpu would have datatypes indicating if it's an integer or not... in case I worry about wasting a single bit for a fricking sign bit lol I could add another datatype which would be unsigned but adding another data type would require even more bits lol. So I am glad you made me think about this... this is a non issue ;) no mixed integer types for me... just a single integer type with a sign bit... unless I need some form of compatibility with other systems ;) woeh.

No inaccurate floating bullshit for me... just rational numbers baby ;)

Bye, Skybuck =D And now I go play dungeon siege 2222222222 yeah. :P

Windows xp has a calculator with a button called "log"

This is actually a Log10 button.

So I use log10 to avoid any confusion.

However you might be right that the implementation of Log2 is much simpler than Log10. Log2 might be closer to it's root so to speak ;). I don't understand logarithm that well... it could have something to do with the number 2 lol :) and root stuff etc ;) wieeee.

No it's "squared root" at least so says my delphi manual.

So "squared root" 2 is 1.414

Hmmm how is that proven ? ;)

selection.

I read in sci.electronics.design that Skybuck Flying wrote (in ) about 'The variable bit cpu', on Sat, 20 Aug 2005:

The correct words are 'the square root of 2'. Not 'squared'. Especially because the phrase 'root 2' is interpreted as 'square root of 2', which is why I understood your 'squared root of 2' to mean 1.414...^2 = 2.

It involves some math far too advanced for me. But I am assured that it is so.

You certainly are right it isn't about compression. Its about fluffing things up with a lot of needless bits.

I was suggesing that you make it handle signed values not complex numbers. Doing signed values does not add any hardware to the ALU. The only place it usually adds a bit is to the status (condition code) register.

I suggest you think about multiply and divide for a bit.

That will be even harder.

what

is

bandwith?

Wax your dick, so it fits better into your donkey's ass lol. hahahaha

Bye, Skybuck.

bandwith?

Which are still inflexible at the instruction level ohhhhhhhhhhh =D ;)

Why would I want to get my dick anywhere near you? I must say, you do have your fantasies!