[KLUG Members] Re: Yo!

[Bryan J. Smith]

Quoting "Robert G. Brown" <bob@acm.org>:
> My "new" box is a 233 mHz Pentium MMX, so anything I can do to snap up
> response will be a welcome thing. If GNOME gets too tedious, I'll be
> able to drop back to one of the lightweght WM's.

The nice thing about Linux is that the "Pentium DE-optimizations" common in
Windows binaries are far less common.  I find it ludicrious that a company calls
a "hack" that makes a program run 1/5th the speed on a 386/486, because it
didn't implement something else as good as a 486, an "optimization."  The
Pentium's ALU had so many flaws in design.

AMD calls such things "errata."  ;->

> Interesting. Perhaps this is an issue with the video card and X-Server?
> I'm sure you're right aboutthe image processing issues, I just wonder
> how much of this we can chalk up to a bad video card, as opposed to a
> bad CPU.

Sounds like memory bandwidth issues from the symtopms he's having.

Also, some early, non-Intel chipsets had serious AGP performance issues.  Intel
didn't detail the AGP spec, and even _changed_ the final specs (_on-purpose_),
right down to the fact that some video card manufacturers first AGP products
were off 50 mils in spacing, mechanically.  Thus, I can only imagine the
electrical issues that ViA, ALi, SiS and others were scrambling to address.

On a more current note, SiS is a full Intel P4 licensee, and now has access to
all the specs.  They, of course, use AMD's HyperTransport for their system
interconnect in even their Intel chipsets.  ;->

I like SiS for Linux because I've never had an issue with their I2C logic and
ATA controllers.  Even some newer Intel ICH southbridges have given me fits with
IBM and WD drives, although newer ICH2 southbridges seem to be fine now.

> My CPU is slower, but it is Intel; I don't recall the video card,
> but it's not a terrible one... we'll see once I plop it in place
> and muck with it a little.

Yes, place an AMD in the same board, see what you get.  Although there are some
K6 optimizations that most older mainboards don't support, you still should get
within 10-15% of actual.

The problem with AMD isn't AMD, but the mainboard manufacturers and OEMs.  E.g.,
Gateway 2000 paid top dollar for their Intel-approved/recommended components,
but when it came to their AMD systems, they didn't even use AMD-approved
mainboards, let alone adequate power supplies (they used cheaper ones that Intel
wouldn't even let them use in their Intel sysetms!) all in the name of "bottom
dollar."

Luckily, this has drastically improved in the Socket-462 age of the Athlon.  And
the new Socket-754/940 basically removes a lot of issues, because most of the
northbridge is now in the Athlon-64/Opteron chips themselves.

-- 
Bryan J. Smith, E.I.             Contact Info:  http://thebs.org
A+/i-Net+/Linux+/Network+/Server+  CCNA CIWA CNA SCSA/SCWSE/SCNA
----------------------------------------------------------------
The reason why Microsoft is so successful is because they fooled
an overwhelming majority of people into thinking there is such a
thing as a "free lunch."  The DOJ case has arisen from those
same consumers who won't accept the fact that they were respon-
sible for locking themselves into only one company's products.