From: | Fade <runefo@***.UIO.NO> |
---|---|
Subject: | Re: Real-Life Computing ... |
Date: | Thu, 28 May 1998 14:16:10 +0000 |
This is on schedule with expected speed growth - about 60% P.A.
That means other server/heavy duty chips will be in the 2.5GHZ
range, same with experimental versions of the same chip.
(Intel ran a pentium chip at 600MHZ two years ago, water cooled.).
Memory and I/O is allready a bigger bottleneck to computing speed
than the chip's MHZ rating, and it increases at a slovenly 7% P.A.
(As Ereskanti pointed out, it's not the chips that keeps the Matrix
in the future, it's the WAN's.. and interface, of course.).
As for the 2K 'bug' (It's an undocumented feature! :) it should not
be a problem for any newer applications - an easy way to check wether
you'll be affected is to play around with the system clock. The most
likely errors is for programs where you supply only the last two
digits of the year - 98, 99, etc - and considers 00 as 'earlier' than
99. It's not system - halting errors, and should be possible to work
around, but it might be annoying and lead to extra work... ). Most of
the system problems should be fixable by a patch - database updating
might be a problem, depending on who made the base.
That's my take on it.
--
Fade
And the Prince of Lies said:
"To reign is worth ambition, though in Hell:
Better to reign in hell than to serve in heaven."
-John Milton, Paradise Lost