Why not skip 64 bit and go right to 128+?

Beattie

Golden Member
Sep 6, 2001
1,774
0
0
You know that in like 5 years everyone is going to be re-doing all the programs and stiff all over again just to support the new 128 bit processors. Why is so much effort going into 64 bit when they should be thinking ahead to bigger and better things?
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
because its taken 10 years just to get close to 32 bit computing. And windows still isn't completely done.
 

Bignate603

Lifer
Sep 5, 2000
13,897
1
0
Complexity of the processors goes up exponentially. It would increase the amount of time to design them, and increase the price of them by quite a bit because of the amount of real estate they would take up. It's not that simple to make these things...
 

Eli

Super Moderator | Elite Member
Oct 9, 1999
50,419
8
81
Because that's not how it works. :p

8 > 16 > 32 > 128? :p
 

Howard

Lifer
Oct 14, 1999
47,982
11
81
Originally posted by: Eli
Because that's not how it works. :p

8 > 16 > 32 > 128? :p
Since when has

8 been greater than 16?
8 been greater than 32?
8 been greater than 128?
16 been greater than 32?
16 been greater than 128?
32 been greater than 128?
 

bigalt

Golden Member
Oct 12, 2000
1,525
0
0
where are you guys living? i've had a commodore 128 for at least like 5 years.
 

jfall

Diamond Member
Oct 31, 2000
5,975
2
0
That is like saying people will eventually be using 5ghz processors, why not skip from 3ghz to 5
 

bleeb

Lifer
Feb 3, 2000
10,868
0
0
Originally posted by: jfall
That is like saying people will eventually be using 5ghz processors, why not skip from 3ghz to 5

What comes after GHz? TeraHertz??
 

Bootprint

Diamond Member
Jan 11, 2002
9,847
0
0

Which processors are you talking about, general cpu's like the P4 or a graphic's chip?

What is the point of going up from 32bits to 64 any ways?

 

FeathersMcGraw

Diamond Member
Oct 17, 2001
4,041
1
0
Originally posted by: Beattie
You know that in like 5 years everyone is going to be re-doing all the programs and stiff all over again just to support the new 128 bit processors. Why is so much effort going into 64 bit when they should be thinking ahead to bigger and better things?

Because the people designing the processors understand things like cost/complexity ratios, Moore's Law, and 128-bit address space.
 

Eli

Super Moderator | Elite Member
Oct 9, 1999
50,419
8
81
Originally posted by: Howard
Originally posted by: Eli
Because that's not how it works. :p

8 > 16 > 32 > 128? :p
Since when has

8 been greater than 16?
8 been greater than 32?
8 been greater than 128?
16 been greater than 32?
16 been greater than 128?
32 been greater than 128?

Bah.. that's not what I meant... they're just arrows.

8 -> 16 -> 32 -> 128

Better? :p

Originally posted by: bleeb
Originally posted by: jfall
That is like saying people will eventually be using 5ghz processors, why not skip from 3ghz to 5

What comes after GHz? TeraHertz??

Yes.

MegaHertz, GigaHertz, TeraHertz, PetaHertz, ExaHertz...
 

XZeroII

Lifer
Jun 30, 2001
12,572
0
0
Why not just skip straight over to quantum computing? Or using neural nets?

Actual answer: because there are VERY few applications that will exceed the need for 64bits. The highest number that a 32bit processor can work with w/o software is 4,294,967,296. There are more people in the world than that. We need something bigger. A 64 bit comptuer can handle numbers as big as 1.84467440737e17. I don't think we will be needing numbers that big in our PC's for quite awhile.
 

hans007

Lifer
Feb 1, 2000
20,212
18
81
Originally posted by: Beattie
You know that in like 5 years everyone is going to be re-doing all the programs and stiff all over again just to support the new 128 bit processors. Why is so much effort going into 64 bit when they should be thinking ahead to bigger and better things?

when you design a CPU all the functional units would have to be twice as wide.



i.e. to make a 128bit alu basically means gluing 2 64bit ones together. that would make CPUs twice the size, for no reason since 128bits is way more than we need, even 64bit probably is. we dont need 128bit precision, nor do we need 128bit addressing. fpu's already have 80bit precision as it is, i'm pretty sure. so yes, we dont need it.

128bit adders, 128bit shifters, etc. not good at all.


a better idea, would be to destroy the x86 isa and start over with one that wasnt designed to save money and resources when everything was super expensive.

greater than 64bit numbers can be respresented using data structures in a 32bit cpu even. it will just take many passes to crunch it through. but still not really a big deal, its not that often used.