| Dux Computer Newsletter
August 16, 2001; Vol. 1, No. 9
IN THIS NEWSLETTER
Recent News and Commentary
Around the Internet
Recent News and Commentary. Various news media
have been expecting Microsoft to release version 6.0 of the Internet Explorer
(IE) this week. I have not seen it yet on the IE
page of Microsoft's web site. This is the same version of the browser
the will be in the forthcoming Windows XP, which is scheduled for release in
October, but may start to appear on new computers as early as next month.
I think Gates has won the browser war unless the U.S. Government
et al bring him to his knees. I have read that 90% of the browsers in
use out there are now IE. Frankly, I hope he does OR some organization
produces a universally used, standard browser (or browsers) with consistently
predictable results, because designing web pages to comply with several different
browser implementations is a royal pain in the neck and a huge waste of time. IE
6.0 will most likely be THE standard. I don't think webmasters
will be doing many compatibility checks after it is in widespread use.
VIA Technology announced yesterday
that they have started volume shipments of their P4X266 motherboard chipset
for the Intel Pentium 4 (P4) processor. Although SiS is trying to play
catch-up and has announced that
it will start sampling their SiS645/SiS961 P4, DDR333 chipset this month, the
VIA chipset is the only one, including those from Intel, that will soon be
widely available in shipping motherboards that will support both the P4 and
Double Data Rate (DDR) Memory.
Intel currently has two chipsets that support the P4: the
i850 and the i845. The i850 marries the processor to RDRAM memory
and the i845 marries it to Single Data Rate SDRAM (PC133 and PC166) memory,
the most common kind of memory in use today. As shown by AnandTech
benchmarks, the first combination is faster than the P4 with the
VIA chipset and DDR memory and the second combination is slower. Because
VIA does not have a license from Intel for the P4 bus interface for its P4X266
chipset and is not charging more for the chipset to pay for the license, and
RDRAM is noticeably more expensive than DDR memory, while DDR memory is quickly
becoming cost competitive with SDRAM, the P4X266 will enjoy a price/performance/MHz
edge over the other two combinations. At the same time DDR motherboards
for the AMD Athlon and Pentium III processors are now coming on line and ramping-up. All
of this will drive DDR memory into the mainstream at an accelerated pace.
Meanwhile, Intel has been rattling swords implying that
it may take VIA to court over the licensing issue. VIA says it doesn't
need a license because of an arrangement with S3, which it says has a license
(VIA bought S3's graphics business some time ago and what remains of S3 is
now called SONICblue).
So, what's likely to happen? VIA has been in licensing
wrangles with Intel before and settled out-of-court. I don't know the
particulars. It seems to me that it would be akin shooting oneself in
the foot as far as processor competition with AMD is concerned if Intel proceeded
with vigor against VIA before there other DDR P4 chipsets and motherboards
to supply the mainstream PC market. Therefore, it probably will not happen,
if at all, until well into next year when Intel should be shipping its own
P4 DDR chipset in quantity.
What does this mean? Intel is stuck between a RamBus
and an Athlon; ALi does
not appear to be in the race yet; SiS is behind the curve and will add a significant
premium, as will ALi, to the cost of its chipset to pay licensing fees to Intel;
and VIA has almost a free reign for the next generation economical DDR chipsets
for P4 motherboards used in mainstream computers.
That is, if VIA can convince first tier motherboard manufacturers
to use the chipset. Currently, they appear reluctant to do so because
of the Intel threats of legal action. Interestingly, today, VIA stated
that they cannot find an Intel patent for the P4 bus. Is there such a
The processor MHz war continues and AMD is losing. Intel
has announced that
it will introduce a 2 GHz version of its Pentium 4 processor on the 27th of
this month. Benchmarks have clearly shown that the 1.4 Athlon is faster
in many respects than a 1.7 GHz P4; it is a more powerful processor. Well,
you can play the more powerful processor just as long as it is the more powerful
processor. A slower architecture clocked at faster and faster speeds
will sooner or later overtake a superior architecture that is not. I
think 2 GHz is that point. It will make the P4 more powerful than the
fastest thing that AMD has to offer.
"MHz is a myth." Back a few years ago Cyrix
had a more powerful processor (so they said) running at a slower clock speed. Well,
it was and it wasn't. As I recall, it did well when running business
benchmarks and worse than either Intel or AMD processors when performing floating-point
operations (the kind of operations that make games and other number crunching
operations fast). So, Cyrix would take their 187.5 MHz 686MX processor,
for example, and give it a performance rating of PR233. That meant that
it was supposed to perform as fast as a 233 MHz Pentium, which it didn't in
all scenarios, especially when playing games and synthesizing music. The
Athlon does not have these weaknesses. Nevertheless, I think the Athlon's
reputation for performance will be sullied if AMD pursues, as they appear to
be getting ready to do, an advertising campaign that claims, "MHz is a myth" when
it isn't entirely a myth. "MHz" do make a difference.
the news section our web site for more news.
Tech Tip. What is the maximum length of a printer cable? There
are two kinds widely used parallel port printer cables for PCs. The old "standard" IBM
Printer cable and the newer IEEE 1284 bi-directional cable commonly sold today. The
old IBM printer cable was a rather loose standard having its roots with cables
originally designed for Centronics printers. It was uni-directional from
the PC to the printer and had a 25-pin male DB25 connector at one end and
a 36-conductor male Centronics connector at the other end. The "standard-issue" cable
was six feet. The maximum cable length was commonly stated as 15 feet. I
have operated printers with old 20-foot IBM PC cables and some, not many, PC/printer
combinations will work with a quality 25-foot cable. There are devices
that can extend this length significantly.
Commonly available IEEE 1284 cables have a 1284 Type A, 25-pin male DB25 connector
at one end and either the older 1284 Type B, 36-conducor male Centronics
connector or the newer 1284 Type C, 36-conductor Mini-Centronics connector
at the other end and are available in 6, 10, 15,
20, and 30 foot lengths. The specification also calls for various cable
configurations with different connector genders. The maximum specified
length of a IEEE 1284 cable cable is 10 Meters (approx. 30 feet) for
data transfers at rates up to 2 MHz (of course, this maximum requires that
the ports at both ends of the cable meet the IEEE 1284 spec as well as the
cable). IEEE 1284 cables should work with older IBM printer ports (SPP
= Standard Printer Port) and uni-directional printers. Answer: 30
Have a computer or networking problem? Visit
our forums for free advice. Our forums are unlike many on
the Internet in that our readers and I try hard to answer almost everyone
who posts a question.
Around the Internet. TCP port numbers have to be
configured in a firewall or router before many Internet games, chat and conferencing
services, and other applications will work. Although one can find many
lists of port numbers on the Internet, the most official list and the one that
appears to be maintained more frequently is available at Internet Assigned
Names Authority's web site here. Go
to FirstGov to
track your Social Security benefits, apply for a federal student loan,
find the nearest veterans hospital, reserve a campsite at a national park...
27 million Web pages.
To view back issues click here. If
you need to E-Mail us regarding the Newsletter, click here.
Disclaimer, and Trademark Information Copyright © 1997-2001
Larry F. Byard. All rights reserved. Except as stated below,
this material or parts thereof may
not be published, broadcast, rewritten, or redistributed by any means whatsoever
without explicit, written permission from the author.
You may circulate
copies of the Dux Computer Newsletter by MANUALLY forwarding it, providing
you forward the issue in its entirety, no fee is involved, and you forward
no more than three issues to any one individual. You may not attach
advertising or otherwise modify the text of the newsletter.