VAX Architecture (was KA620)

From: Paul Koning <pkoning_at_equallogic.com>
Date: Sat Mar 20 14:28:53 2004

>>>>> "Michael" == Michael Sokolov <msokolov_at_ivan.harhan.org> writes:

 Michael> OK, I've already figured myself that all the really
 Michael> important "what is a VAX by definition and how to build one"
 Michael> stuff *is* in the published VARMs, so my project of building
 Michael> a new VAX is not stuck waiting for KGB to pry STD 032 out of
 Michael> DEC. I fully understand the VAX Architecture (and have a
 Michael> solid rigorous spec definition) based on the 3 VARMs I have

One nice thing about DEC is that it took rigorous architecture
definition very seriously starting with the VAX (and, earlier still,
in networking). There aren't many other companies where that is true
-- that you can take the spec and build what it says and have it be
correct and complete.

By the way, if you have the right connections around Russia, you could
track down the people who built the VAX clones back in the days of the
USSR. I believe STIMTI in Vilnius was involved...

>> I think I've said before, what you really want is not STD 032 but
>> AXE, the tool that runs on your new VAX and checks for correct
>> operation of instructions.

 Michael> I've never heard about AXE from you, but I have heard about
 Michael> it from other sources. Yes, that would really be nice.

AXE was the tool that DEC used to verify that a new VAX had the CPU
done completely right. They found that a random test like that, run
long enough, was far more rigorous than any conventional diagnostic.

 Michael> Interestingly, however, it appears that at some point there
 Michael> were diagnostic programs available to the general public
 Michael> that, judging from the descriptions, apparently do similar
 Michael> instruction testing, though they were presumably intended
 Michael> for troubleshooting broken hardware rather than for
 Michael> validating new implementations. The KA820 Technical Manual,
 Michael> for example, refers to these: ...
 Michael> The descriptions of these diagnostics sound very much like
 Michael> AXE. Any idea where to find these diags?

AXE isn't a diagnostic, it's an implementation verifier. The others
definitely are diagnostics, intended to find specific problems in the
field quickly. AXE you run for weeks, and if it finds a problem you
have to do a bunch of digging to find out what the problem details
are.

        paul
Received on Sat Mar 20 2004 - 14:28:53 GMT