🚀 go-pugleaf

In article <12867@sail.LABS.TEK.COM> arnief@sail.LABS.TEK.COM (Arnie Frisch) writes:
>>Do you realize that the DN10000 scan architecture was developed before
>>1149.1?  How can you accuse us of "departing from the standard" when
>>the standard wasn't even around when we did this work?
>
>Do you realize that JTAG (the Joint Test Action Group) had been working
>on this for about 8 years, and that they were well ahead of the DN10000
>architecture when they started writing the basis for the P1149.1
>standard?

The DN10000 architecture work began in 1985, and we had a working
system in 1987.  This included power-up self tests that utilized
pseudorandom scan test techniques, an offline scan test program that
tested the system with automatically-generated scan patterns, and a
remote debugger that allowed access to the system under test from any
other workstation on the net.  We never would have finished any of
this if we had waited for 1149.1; the market window for the DN10k
would have closed by then.

What do you mean when you say that the 1149.1 folks were "well ahead"
of us?  Especially given the fact that we completed our work before
they published the standard?  Just because they were working on 1149.1
for a portion of the time that we were working on the DN10k doesn't
mean that their work was anywhere near being ready to be put to use in
a real system.  Be practical, Arnie.

>As I understand it, consultants were hired to do the major part of the
>IC design for the 10K - including the scan design.  They subsequently
>came out with a product for testing scan designs and tried to sell it -
>unsuccessfully - under the banner "Gillytronics".  Seems like something
>was lacking in their methodology.

Sorry, Arnie.  You're talking to a member of the DN10k development
team, so please get your facts straight.  We did our own ICs using
gate arrays from Integrated CMOS Systems (ICS).  The scan architecture
was largely the brainchild of Bulent Dervisoglu.  His many papers and
articles contain a lot of good information related to scan
architectures and scan testing.  (Judging from the content of your
postings, you should read them sometime.)  Gillytronics was somehow
related to ICS, but that is their only (weak) link to the DN10k work.
Remember, Arnie, they set out to do chip testing with a specialized
test system, and we were doing system testing using a built-in service
processor.  Very different beasties altogether.

Why are you so bent out of shape about this, Arnie?  Seems like you
need to broaden your horizons a bit.

-steve