urip@nsc.UUCP (05/23/86)
I was quite amazed at the amount of dis-information that is running around
about this subject. I have been with National Semi' for almost 5 years now
(yes, five!!), so I KNOW the TRUTH. In fact, I started my way at NSC in the
team that designed the DMA chip (was 16203), but I didn't like being
a chip designer so I moved to the software group. Anyway, I know
personally the people who designed the CPU and the rest of 32***.
I am new here on the net (why did I wait so long? read to the end and see...),
so I want first to clear the cloud of rumors and twisted facts about the
names of the chips.
You will probably hear from me in the future about other popular subjects
like 32332, optimizing compilers for 32000 etc.
But now, for some history.
In the beginning, there were no 32 bit u-processors, only 16 bit.
I'm talking late 70's (at that time I didn't even know how a computer
looked like on the inside). 32 bit uP's did not seem to be hot on the market,
so marketing called the family "NS 16000". There were two versions of the
CPU: 16032 and 16008. Both CPU's were internally identical, using 32 bit
internal registers and buses, but they differed in their external buses:
16 and 8 bits. Later, the 16032 was improved to use a 32 bit external bus,
and this version was named 32032. Now it became clear the something is
wrong in the naming convention. We have three versions of CPU, all of them
are 32 bit internally, and the externally:
16008 - 8 bit ext. bus
16032 - 16 bit ext. bus
32032 - 32 bit ext. bus
so in the 16008, the right digits stand for the bus-width,
in the 16032 the left digits stand for the bus-width and
in the 32032 it is not clear at all who stands for what.
Above all, the market has changed (it was clear that 32 bit uP is THE thing)
and the marketing people have also changed!
There were long debates within NSC about changing the names. In addition
to the chips we also had a set of software tools called:
PAS16, ASM16, DBG16 ...
Should these names change too?
The decision was made. I think it was a good decision.
The second source agreement with TI was a good occasion for announcing
the new name "Series 32000" (not because of any pressure from TI...).
The legend about the "lost tribe of 16 bit uP's" whose evidence can be
found in the 16 bit limitation of the CXPD instruction is totally untrue,
as well as the description of those limitations (the limitations themselves
are a subject to another article, if anybody is interested).
For the same reason that I thought it was a good decision to change the
name "Series 32000" to "NS 16000", I am very much in favor of changing
the name of this group to "ns32k" (not "ti32k" because TI did not design
the chips, only second source them).
On my first scan of news groups I skipped micro.16k, so I am sure that
EVERYONE does the same mistake. I was lucky to have good friends to
correct me, but think how many net readers were not that lucky!
I haven't answered yet why it took me so long to go on the net.
You may have also wondered why, as a software person, I haven't used any
slang words and idioms.
Well, the reason for both is that I work at NSTA (National Semi' Tel Aviv).
So first, we don't get net-news there (yet), and second, my native language
is Hebrew and not English. You probably won't be able to pronounce most of
the names of the 32000 designers (or my name) but this discussion belongs
already to net.nlang.
Uri.
--
Uri Postavsky - National Semiconductor Tel Aviv
Until end of June my path is:
{cbosgd,fortune,hplabs,inhp4,seismo}!nsc!uripchongo@nsc.UUCP (Landon Noll) (05/24/86)
In article <3598@nsc.UUCP> urip@nsc.UUCP (Uri Postavsky NSTA) writes: >I was quite amazed at the amount of dis-information that is running around >about this subject. I have been with National Semi' for almost 5 years now >(yes, five!!), so I KNOW the TRUTH. Your statments were correct, but there is even more to the "truth" than you stated. Allow my 3 years at NSC to denote: 16016 (never made) 32132 (now 32532) 32032E (now 32332) 16201 (now 32201) 16202 (now 32202) 16203 (now 32203) which were planned but never made. And let us not forget the: 8096 which is an important number in the early days of the 32000. chongo <such is the numbers game> /\32/\