end of moore's law?

Adul

Elite Member
Oct 9, 1999
32,999
44
91
danny.tangtam.com


<< . By comparison, the transistors found in the latest chips in use today measure 0.18 microns from one side of the transistor gate to the other. >>



geforce 3 is at .15
 

nortexoid

Diamond Member
May 1, 2000
4,096
0
0
latest cyrix chips use .15 micron...and i believe there's one other chip that uses an &quot;integrated&quot; .15+.18 technology...or is that .13+.18...whatever...

anyway, whenever the end comes, computers will probably be as fast as we could ever want them (if that's possible)
 

MGMorden

Diamond Member
Jul 4, 2000
3,348
0
76
There will always be room to push forward. These guys are predicting &quot;the end&quot; with our current scientific knowledge? I'm sure 400 years ago scientists would have thought a processor as we know it now was completely impossible. One of these magic boxes that does things on it's own just isn't possible. Our predictions will grow such as baseless as time passes on. Even if for a while we get to a point where we can't get any faster, that doesn't mean it can't be done.



<< is there a real use for 20GHz? >>



For our current apps, no, but for future apps, you bet. At risk of sounding nerdy, think of the holodeck programs on Star Trek. If we were ever able to get a holographic matrix working, what type of processing power do you think it would take to run such a simulation? I'd wager a lot more than 20Ghz.
 

JackBurton

Lifer
Jul 18, 2000
15,993
14
81
forget the holodeck, some of these MAME games take a BUTT load of CPU cycles! At 866MHz, on some games feels like I have a P90!! :|
 

Quickfingerz

Diamond Member
Jan 18, 2000
3,176
0
0
It's not all about the clock cycle. There are 300 Mhz RISC CPU's that can easily double the processing of a 1.7 Ghz P4.
 

frizzlefry

Golden Member
May 14, 2001
1,711
0
0
Right! Think about those SGI Indy machines that were renedering in real time 6 years ago? Those use RISC and they kick butt.

BTW: Is $350 for a 5-6 year old Indy with 17&quot; monitor a good deal? Saw one at the computer show today and I told the guy I'd think about it. He's got a bunch of stuff like SGI servers, Ciscor rackmounts etc.
 

nortexoid

Diamond Member
May 1, 2000
4,096
0
0
aren't internal clock speeds going to be irrelevant anyway in the future?..

with &quot;asynchronous&quot; architectures?...different parts of the chip running at different speeds, etc...

YA KNOW?
 

zzzz

Diamond Member
Sep 1, 2000
5,498
1
76
The current processes for producing smaller silicon transistors is getting to its limits. Intel is still trying to improve the current processing.
Ultimately there will be a fundamental (as opposed to technical) limit to how small you can go as electrons won't behave the same way when they are confined to very small volumes. (ultimate limit being wires of atomic width)
 

ElFenix

Elite Member
Super Moderator
Mar 20, 2000
102,405
8,582
126


<< By the time Intel -- and others -- roll out semiconductors with transistor gates 0.02 microns wide, those chips will last for one more processor generation. Such a generation, in Intel's case, typically lasts about three years. This means that Moore's Law -- formulated by Intel co-founder Gordon Moore decades ago -- will last into the next decade. >>



actually, it looks like transistor sizes last about 2 years in intel's case... maybe less. cumine came out late '99, .18 is getting replaced later this year. deschutes came out early '98, .25 got replaced about 1.6 years later. wonder when .35 came out?
 

ST4RCUTTER

Platinum Member
Feb 13, 2001
2,841
0
0
EUV lithography should take us to the limits of silicon based chips but after that it's anyone's guess. As zzzz hinted, when you start forcing electrons through ever smaller conductors the principles behind these electronic devices tend to break down and quantum mechanics becomes the dominant guiding force. I'm kinda excited though. After all, necessity is the mother of invention...