microcode, compilers, and supercomputer architecture

From: Allison J Parent <allisonp_at_world.std.com>
Date: Tue Apr 6 22:00:49 1999

<> To put this slightly on topic, we all know that computer systems have
<> become more reliable in the last 20 years but does anyone have any real

Potentially they are. there are fewer chips and connectors to fail and
chips can be highly reliable however...

The average PC is tossed togeher from a motly assortment of parts and
software and usually works, sometimes. With that cases are more a fashon
statment rather than given the thought toward cooling... those 333mhz
celerons do get mightly hot and those cheap bushing fans do fail often.

When new system have the uptime records of my 12 year old uVAXII we can
debate this... the record is 422 days contuinious uptime without reboot, it
was limited by a power failure.

I've built Z80 based systems with uptimes measured in years (still
counting as they are battery backed). They are running very debugged code
and power failure is not seen due to power systems that expect brownouts.

It's not to say they cannot out do that, only that reliability is a
measure of quality and sadly most PCs greatly lack the latter.

Technology does not automagically beget reliability, it affords the
opportunity to create it. Often that boat never docked at the airport.

Allison
Received on Tue Apr 06 1999 - 22:00:49 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:31:40 BST