Byte size
pdb...@bnr-vpa.uucp
it just frequently happens to be that size. First some history.
The infamous IBM 7030 (late 1950's, early 60's), also known as
"Stretch", was IBM's first attempt at a supercomputer. Among
other spiffy features, it allowed the programmer to specify the size
of the operands, so if he wanted to do 23 bit arithmetic, he set the
"byte size" field of the instruction to 23. The designers coined the term
"byte" to mean "a bunch of bits operated on by an instruction".
The 360 changed the meaning to 8 bits. Machines with 36 bits, such as
the PDP-10, Honeywell mainframes, and Univac (sorry, Sperry. uh,
what are they called today?) have either 6 or 9 bit bytes, and the
BBN C-70 has a ten bit byte. Further, the information coding people
talk about "bytes" to mean "a few bits", and talk quite freely about
byte sizes. Finally, the IEEE dictionary defines a byte asa small
number of bits which can be operated on by a machine.
-peter
wsie...@rmi.uucp
wsie...@rmi.uucp
>it just frequently happens to be that size.
A very easy definition of a byte is as follows:
A byte is the smallest addressable unit of memory in a given machine.
This should do it ??
===============================================================
Cabbage: A familiar kitchen-garden vegetable about as large and
as wise as a man's head. (A. Bierce)
===============================================================
Richard Mateosian
I know of one machine on which bits are addressable, but which usefully
maintains the concept of an 8-bit byte as another addressable entity.
Richard Mateosian ...ucbvax!ucbiris!srm 2919 Forest Avenue
415/540-7745 s...@iris.Berkeley.EDU Berkeley, CA 94705
Craig Jackson
>>/***** rmi:comp.arch / bnr-vpa!pdbain / 1:59 pm Dec 5, 1986*/
>>Contrary to popular belief, a byte is not necessarily 8 bits -
>>it just frequently happens to be that size.
>
>A very easy definition of a byte is as follows:
>
> A byte is the smallest addressable unit of memory in a given machine.
>
>This should do it ??
Most of the machines with byte sizes of other than eight bits are also
word-addressable machines. The idea that address bits could be wasted on
addressing to the character came along about the same time as eight-bit
characters. Before that, most machines were word addressable (aside from
a few like the IBM 1401). Some of those machines had useful subdivisions
of the word, and those were called bytes. (In my case, the CDC 6400 had
60 bit words, but its peripheral processors only had 12-bit words. Since
the operating system resided in the PPs, there were many 12-bit fields in
memory (packed five to a word). These fields were typically called bytes.
Incidently, each such byte could hold two six-bit characters.
The definitions I've always worked with are:
A byte is a collection of two or more bits, smaller than a word.
A word is the 'natural' unit of operations on a machine.
(Which means that 32 bits on a vax should really be called a word, not
a longword.)
--
Craig Jackson
UUCP: {harvard!axiom,linus!axiom,ll-xn}!drilex!dricej
BIX: cjackson
Henry Spencer
> (Which means that 32 bits on a vax should really be called a word, not
> a longword.)
The only reason it's called a longword is because of all the semi-subliminal
it's-really-just-a-big-pdp11-so-don't-worry-about-incompatibility marketing
horseshit that surrounded the vax in its early days. Remember when a 780's
official designation was a "VAX-11/780"?
(Well, to be honest, there may be another reason: it may be a historical
survival from the early vax development days, at which time I suspect the
vax was going to be much more of a stretched 11 than a new machine. Note
that "VAX" stands for Virtual Address Extended. Extended what? Extended
pdp11.)
--
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,decvax,pyramid}!utzoo!henry
Mark Campbell
>> A word is the 'natural' unit of operations on a machine.
>> (Which means that 32 bits on a vax should really be called a word, not
>> a longword.)
>
>The only reason it's called a longword is because of all the semi-subliminal
>it's-really-just-a-big-pdp11-so-don't-worry-about-incompatibility marketing
>horseshit that surrounded the vax in its early days. Remember when a 780's
>official designation was a "VAX-11/780"?
> Henry Spencer @ U of Toronto Zoology
> {allegra,ihnp4,decvax,pyramid}!utzoo!henry
[For what it's worth...]
A DEC field representitive in the late 70's told me that the reason that the
VAX word was 16 bits was that the Government classified machines by word
size when purchasing -- thus the VAX would have competed with a higher
class of machines in contracts.
--
Mark Campbell
{}!ncsu!ncrcae!sauron!campbell
Amos Shapir
>The only reason it's called a longword is because of all the semi-subliminal
>it's-really-just-a-big-pdp11-so-don't-worry-about-incompatibility marketing
>horseshit that surrounded the vax in its early days. Remember when a 780's
>official designation was a "VAX-11/780"?
Actually, DEC had made a brilliant (may be unintentional) upgrade,
changing architecture while keeping the customers, doing it both radically
and gradually at the same time. There are very few cases (that I know of)
in which an architecture upgrade had managed to keep all the good parts
and leave out all the bad parts - the hardest part is convincing the users
that it won't hurt; providing 'compatibility mode' is an excellent cushion,
and, unlike other vendors' upgrades, you dont need to know about it
if you dont use it.
--
Amos Shapir
National Semiconductor (Israel)
6 Maskit st. P.O.B. 3007, Herzlia 46104, Israel
(011-972) 52-522261 amos%nsta@nsc 34.48'E 32.10'N
Dave Spain
I'm not sure this ever changed. There (believe it or not) is a good
deal of logic behind this designation and I don't just mean in the
philosophical sense. All VAX processors that I was familar with that used
the "-11/" nmemonic offered some form of hardware assisted compatibility-mode,
typically at the instruction set level.
It wasn't until after a descision was made to offer PDP-11 compatibility
as a software option was the "-11/" convention dropped.
Bjorn R. Bjornsson
> In article <74...@utzoo.UUCP> he...@utzoo.UUCP (Henry Spencer) writes:
> >Remember when a 780's
> >official designation was a "VAX-11/780"?
>
> Actually, DEC had made a brilliant (may be unintentional) upgrade,
It could have been brilliant, but was greatly marred by
flaws like the following:
RTI and RTT ignore the mode bits in the PSW image
on the stack, and set the mode to user without
trapping.
If RTI / RTT had trapped mode changes, or simply been left
unimplemented, you could have written or ported real PDP-11
operating systems to run on the VAX.
As it turned out, the AME (pseudo RSX-11M on a VAX), was
useless except for trivial RSX applications, this because
they left out socalled PLAS directives. What it did do
was allow DEC to push the VAX through the door without any
software. My suspicion has always been that DEC included
compatibility mode in the early VAXen for this reason and
this reason only. After the fashion of computer manufacturers,
you then proceed advertise it as a feature. When you take
it away, you can again advertise that as a feature, this
time saying it's not taking up any resources, and users
don't have to pay for something they don't need or want.
Remember VMS versions 1 through 2. Almost utilities and compilers
ran in compatibility mode. Yeah, DCL translated COPY commands
to RSX PIP invocations, ditto for RENAME, DELETE, ad nauseum.
The Fortran compiler was a modestly hacked version of PDP-11
Fortran-IV-Plus, and on and on and on. After version 2, more
and more things started showing up in native mode. Until
finally DEC was in a position to release VAXen without a PDP-11
mode, which is basically when they turfed it.
It's easy to imagine that DEC would have been at least a year
later in introducing the VAX without compatibility mode. Makes
you wonder what the world would look like today, if that had
been the case.
Don't ascribe to brilliance, that which
can be explained by simple expediency,
Bjorn R. Bjornsson
alberta!bjorn
Bjorn R. Bjornsson
> If RTI / RTT had trapped mode changes, or simply been left
> unimplemented, you could have written or ported real PDP-11
> operating systems to run on the VAX.
Come to think of it, this was probably intentional, just so
that PDP-11 operating systems could not be ported to the
VAX. Brilliant!
Bjorn R. Bjornsson
alberta!bjorn
Amos Shapir
>It's easy to imagine that DEC would have been at least a year
>later in introducing the VAX without compatibility mode. Makes
>you wonder what the world would look like today, if that had
>been the case.
This is *exactly* what I meant by 'brilliance'! There were many projects
that were so advanced they came out years too late because it took too
long to write the software for them. The brilliant idea was: 'let's not wait
for native VAX software, PDP11 software should be just fine for the
introduction phase'. It *is* a success, and you can't argue with that.
(Disclaimer: I do not, and never did, work for DEC).
Dave Martindale
>It's easy to imagine that DEC would have been at least a year
>later in introducing the VAX without compatibility mode.
amos%nsta@nsc (Amos Shapir) writes:
>The brilliant idea was: 'let's not wait
>for native VAX software, PDP11 software should be just fine for the
>introduction phase'.
Let's not forget that the early emergence of the VAX line was also
very good for those of us that didn't (and still don't) care about
DEC-supplied software. I doubt that DEC saw this as an advantage,
but they pretty much ignored UNIX* in those days anyway.
* In "those days", UNIX was a noun, not an adjective.
Dan Kary
> >you wonder what the world would look like today, if that had
> >been the case.
> This is *exactly* what I meant by 'brilliance'! There were many projects
> that were so advanced they came out years too late because it took too
> long to write the software for them.
The Apple Macintosh is a good example. It was two years late. If their
delivery had been more timely MS-DOS and intel processors might be as
popular as they deserve to be.
Dan Kary