Author: Tom Kerrigan
Date: 10:29:31 10/25/01
Go up one level in this thread
On October 24, 2001 at 23:17:20, Robert Hyatt wrote: >On October 24, 2001 at 15:23:14, Tom Kerrigan wrote: > >>On October 24, 2001 at 13:35:48, Tony Werten wrote: >> >>>>Although 128 bit microprocessors are probably 20 years away or longer... >>> >>>Why 20 ? The amount of bits seem to have doubled every 10 years until now. >> >>Computer hardware has dramatically outpaced most computer software, hence the >>slowdown in sales. People are happy with their current processor speed and >>memory. Now that you can get 512MB of RAM for $40, or something silly like that, >>most people have way more memory than they know what to do with, and they aren't >>even approaching the limits of 32-bit addressing, much less the 36-bit >>addressing afforded by the P6/P4. Switch to 64 bit addressing and you get 4 >>_billion_ times more memory than people know what to do with. Not very >>compelling. >> >>Making a 128-bit chip would be pretty easy, kind of like how 64-bit chips are >>pretty easy to make (heck, there's one in the Nintendo 64 and Compaq iPAC). But >>even though 64-bit chips are easy to get, PC users aren't trying to switch >>because there's no need... >> >>-Tom > > >Don't get too hung up on the 64 bit address space. That isn't what has >driven the 64 bit world to date. The important point has always been FP >precision and speed. And 64 bit hardware is more accurate than 32 bit >hardware when dealing with FP. And 128 bit would be even better as it >gives larger exponents and fractions at the same speed. I assumed that by 64-bit, we all meant the integer side of the CPU. Of course FP is going to be wider; it always has been. -Tom
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.