Hi Tony!

10 Oct 2016 07:54, from Tony Langdon -> Richard Menedetter:

 TL>> It was an Athalon, if I recall.
 RM>> Really?
 TL> Yep, and I remember in the end having to underclock it slightly to
 TL> improve stability, which did make a big improvement.

Probably one of the first gen cartridge based ones. (AMD had production
issues ... as usual ;)
Because actually the Athlons were those processors that drove Intel nuts.
(anyone remember the recalled 1.13 GHz Pentium 3s ;)
And then it drove them into the Pentium 4 disaster.
It took quit long until they redeemed themselves with the core series ;)

 TL> I had heard lots of good things about the Athalon,

They are spelled without the additional a -> Athlon

 TL> which is why I went down that route, but for my use scenario, it
 TL> didn't live up to expectations.
 TL> It was an OK system, but not as good as I had hoped.

It blew out anything that Intel had at that time of the water.
What was your exact problem?
I had a 800 MHz Thunderbird, and it was exceptionally great.
Only issue I can think of is with the first cartridge based ones, which had
heat and production issues.

 TL> The P4 was another CPU I wasn't a big fan of, for different reasons.
 TL> The P4 system I ran (still have it) was rock solid, and hyperthreading
 TL> did give it a little performance kick, but it sucked power like it was
 TL> going out of fashion.

The P4 was a complete disaster that Intel only did to have something that
can counter the Athlon marketing wise.
All the DEC engineers went to AMD, and the Athlon microarchitecture was 1-2
years ahead of what Intel had at that time.

See also the talk/book of Bob Colwell for that.

 TL> One thing AMD did do right was developing the 64 bit instruction set
 TL> we know and love today. :)

That was only made possible by the exceptional K7 (Athlon) microarchitecture.
We compared P6 (PPro, P2, P3) to K7 (Athlon) at university ...
The age of P6 showed VERY clearly.

 TL> Intel's "clean slate" 65 bit efforts didn't penetrate the
mainstream
 TL> like AMDs did.

Well ... the Itanium was never really meant as a P4 replacement.
It was a monster that had to be able to replace PARISC for HP.

 RM>> The I/O of the Raspberries is awful.
 TL> Yes, that is a known issue, and one I've kept in mind, so don't put a
 TL> Pi into a situation demanding high I/O performance.  It's simply not
 TL> designed for that.

Indeed!

CU, Ricsi

--- GoldED+/LNX