Don't know how many people here watch "Young Sheldon" which is a spinoff of "The Big Bang Theory". Anyway, last night we see Sheldon Cooper take delivery of a main frame in his dorm. What got rolled in was a VAX-6000-420 with three racks of disk storage. I guess Chuck Lorre and staff didn't know this was a mini (or perhaps they did but just liked the sound of Sheldon say "I need a mainframe")

Neil Rieck
Waterloo, Ontario, Canada.
http://neilrieck.net
http://neilrieck.net/OpenVMS.html

On 2/3/2023 8:48 AM, Neil Rieck wrote:
> Don't know how many people here watch "Young Sheldon" which is a
> spinoff of "The Big Bang Theory". Anyway, last night we see Sheldon
> Cooper take delivery of a main frame in his dorm. What got rolled in
> was a VAX-6000-420 with three racks of disk storage. I guess Chuck
> Lorre and staff didn't know this was a mini (or perhaps they did but
> just liked the sound of Sheldon say "I need a mainframe")

Ah memories.

End of 80's I worked on a 6000-420. We could have 110 interactive
users on that system.

Arne

On 03/02/2023 14:01, Arne Vajhøj wrote:
> On 2/3/2023 8:48 AM, Neil Rieck wrote:
>> Don't know how many people here watch "Young Sheldon" which is a
>> spinoff of "The Big Bang Theory". Anyway, last night we see Sheldon
>> Cooper take delivery of a main frame in his dorm. What got rolled in
>> was a VAX-6000-420 with three racks of disk storage. I guess Chuck
>> Lorre and staff didn't know this was a mini (or perhaps they did but
>> just liked the sound of Sheldon say "I need a mainframe")
>
> Ah memories.
>
> End of 80's I worked on a 6000-420. We could have 110 interactive
> users on that system.

Memories indeed. The 6000-4n0 was the lower-end system to support
vector processors (the other one was, of course, the VAX 9000).

Managed both of these in the early 1990s in Heidelberg at the
European Molecular Biology Laboratory.

P.s. I have no idea who "Sheldon Cooper" is.

Arne Vajhøj <arne@vajhoej.dk> wrote in
news:trjqsr$1io8j$1@dont-email.me:

> On 2/3/2023 3:22 PM, Neil Rieck wrote:
>>
>> https://en.wikipedia.org/wiki/Sheldon_Cooper
>>
>> VAX-6000 was the Chevy of minicomputers in the 1990s.
>>
>> IIRC, Vector processing was only available if you purchased the
>> optional vector processing board (I saw one once at DEC in Bedford
>> Mass.)
>> https://manx-docs.org/collections/mds-199909/cd1/vax/60vaapg1.pdf
>
> Yes - it was HW addon.
>
> I don't think it sold that well. Those RISC thingies was
> starting to pop up everywhere.
>
> But the 6000 series sold pretty well and I think a lot of them
> kept running well up in the 00's. They were pretty
> good systems - especially the 400 and later.
>
> Unlike the 9000's which is think mostly had a short
> career in the data center.
>
> Arne

A friend of mine ran a geo-physical data processing company in Calgary. I
visited his office around 1990 and they had the VAX 11/780 they had
started with still running and they had started with some vector
processing hardware attached to the 780. By then they were using high
powered workstations for the number crunching and the 780 was just used
for reading the raw data from the 9 track tapes the customers sent them.

--
Clark G
* take away the em's to reply

On Wednesday, February 8, 2023 at 1:27:34 PM UTC-8, Bob Eager wrote:
> On Wed, 08 Feb 2023 19:19:30 +0000, Scott Dorsey wrote:
>
> > IBM guys would not have. My boss would have said that if it can't back
> > up the DASDs with the CPU halted than it wasn't a real mainframe.
> I know it was probably later on, but couldn't you do that with MSCP disks?

Yes, the HSC50 had a direct disk-to-tape backup utility that didn't require the CPU. Consistency was your problem so it was mostly used off line. The system operated a TU78 tape drive at absolute max speed. Forward write speed was almost rewind speed. You could always tell when it was running by the sound of the vacuum columns.

On 2/11/2023 11:08 AM, Tony Priborsky wrote:
> On Wednesday, February 8, 2023 at 1:27:34 PM UTC-8, Bob Eager wrote:
>> On Wed, 08 Feb 2023 19:19:30 +0000, Scott Dorsey wrote:
>>
>>> IBM guys would not have. My boss would have said that if it can't
>>> back up the DASDs with the CPU halted than it wasn't a real
>>> mainframe.
>> I know it was probably later on, but couldn't you do that with MSCP
>> disks?
>
> Yes, the HSC50 had a direct disk-to-tape backup utility that didn't
> require the CPU. Consistency was your problem so it was mostly used
> off line. The system operated a TU78 tape drive at absolute max
> speed. Forward write speed was almost rewind speed. You could
> always tell when it was running by the sound of the vacuum columns.

Did it generate a backup in VMS Backup format or did
it just dump physical disk blocks to tape blocks?

Arne

On Saturday, February 11, 2023 at 8:14:16 AM UTC-8, Arne Vajhøj wrote:
> On 2/11/2023 11:08 AM, Tony Priborsky wrote:
> > On Wednesday, February 8, 2023 at 1:27:34 PM UTC-8, Bob Eager wrote:
> >> On Wed, 08 Feb 2023 19:19:30 +0000, Scott Dorsey wrote:
> >>
> >>> IBM guys would not have. My boss would have said that if it can't
> >>> back up the DASDs with the CPU halted than it wasn't a real
> >>> mainframe.
> >> I know it was probably later on, but couldn't you do that with MSCP
> >> disks?
> >
> > Yes, the HSC50 had a direct disk-to-tape backup utility that didn't
> > require the CPU. Consistency was your problem so it was mostly used
> > off line. The system operated a TU78 tape drive at absolute max
> > speed. Forward write speed was almost rewind speed. You could
> > always tell when it was running by the sound of the vacuum columns.
> Did it generate a backup in VMS Backup format or did
> it just dump physical disk blocks to tape blocks?
>
> Arne

I'm pretty sure it didn't write a native VMS backup saveset - remember the CPU on the HSC was a PDP-11. Logically the backup was the equivalent of VMS "BACKUP/PHYSICAL". I seem to recall that at some point the VMS backup could restore (read) a HSC backup.

On 2/11/2023 11:18 AM, Tony Priborsky wrote:
> On Saturday, February 11, 2023 at 8:14:16 AM UTC-8, Arne Vajhøj
> wrote:
>> On 2/11/2023 11:08 AM, Tony Priborsky wrote:
>>> On Wednesday, February 8, 2023 at 1:27:34 PM UTC-8, Bob Eager
>>> wrote:
>>>> On Wed, 08 Feb 2023 19:19:30 +0000, Scott Dorsey wrote:
>>>>
>>>>> IBM guys would not have. My boss would have said that if it
>>>>> can't back up the DASDs with the CPU halted than it wasn't a
>>>>> real mainframe.
>>>> I know it was probably later on, but couldn't you do that with
>>>> MSCP disks?
>>>
>>> Yes, the HSC50 had a direct disk-to-tape backup utility that
>>> didn't require the CPU. Consistency was your problem so it was
>>> mostly used off line. The system operated a TU78 tape drive at
>>> absolute max speed. Forward write speed was almost rewind speed.
>>> You could always tell when it was running by the sound of the
>>> vacuum columns.
>> Did it generate a backup in VMS Backup format or did it just dump
>> physical disk blocks to tape blocks?
>
> I'm pretty sure it didn't write a native VMS backup saveset -
> remember the CPU on the HSC was a PDP-11. Logically the backup was
> the equivalent of VMS "BACKUP/PHYSICAL". I seem to recall that at
> some point the VMS backup could restore (read) a HSC backup.

So no CRC and no redundancy groups?

Arne

On Saturday, February 11, 2023 at 8:26:48 AM UTC-8, Arne Vajhøj wrote:
> So no CRC and no redundancy groups?
>
> Arne

No.

On 2023-02-11 17:26, Arne Vajhøj wrote:
> On 2/11/2023 11:18 AM, Tony Priborsky wrote:
>> I'm pretty sure it didn't write a native VMS backup saveset -
>> remember the CPU on the HSC was a PDP-11.   Logically the backup was
>> the equivalent of VMS "BACKUP/PHYSICAL".  I seem to recall that at
>> some point the VMS backup could restore (read) a HSC backup.
>
> So no CRC and no redundancy groups?

Depends on how you mean that. Tapes have CRC on the data unless I
remember wrong. Just like disks have. But no redundancy or ECC, so if
media got corrupted, you loose.

But HSCs are pretty nice. Been a while since I used one.

Johnny

Johnny Billquist <bqt@softjar.se> writes:

> On 2023-02-11 17:26, Arne Vajhøj wrote:
> > On 2/11/2023 11:18 AM, Tony Priborsky wrote:
> >> I'm pretty sure it didn't write a native VMS backup saveset -
> >> remember the CPU on the HSC was a PDP-11.   Logically the backup was
> >> the equivalent of VMS "BACKUP/PHYSICAL".  I seem to recall that at
> >> some point the VMS backup could restore (read) a HSC backup.
> >
> > So no CRC and no redundancy groups?
>
> Depends on how you mean that. Tapes have CRC on the data unless I
> remember wrong. Just like disks have. But no redundancy or ECC, so if
> media got corrupted, you loose.

Hmm. In 50+ years of data processing/IT/systems work, I've never heard of
tapes having (nonprogrammatic) CRC, that is, anything generated by the
hardware/firmware rather than the backup program.

> But HSCs are pretty nice. Been a while since I used one.
>
> Johnny

--
Rich Alderson news@alderson.users.panix.com
Audendum est, et veritas investiganda; quam etiamsi non assequamur,
omnino tamen proprius, quam nunc sumus, ad eam perveniemus.
--Galen

On 2/11/2023 7:02 PM, Johnny Billquist wrote:
> On 2023-02-11 17:26, Arne Vajhøj wrote:
>> On 2/11/2023 11:18 AM, Tony Priborsky wrote:
>>> I'm pretty sure it didn't write a native VMS backup saveset -
>>> remember the CPU on the HSC was a PDP-11.   Logically the backup was
>>> the equivalent of VMS "BACKUP/PHYSICAL".  I seem to recall that at
>>> some point the VMS backup could restore (read) a HSC backup.
>>
>> So no CRC and no redundancy groups?
>
> Depends on how you mean that. Tapes have CRC on the data unless I
> remember wrong. Just like disks have.

As I remember 9 track tapes then it was pure data blocks
and tape marks - no CRC.

If mounted as labeled tape then some blocks were
considered VOL and HDR blocks, but that was an OS
feature not a drive feature.

Arne

> Hmm. In 50+ years of data processing/IT/systems work, I've never heard of
> tapes having (nonprogrammatic) CRC, that is, anything generated by the
> hardware/firmware rather than the backup program.

Seven track tape had both frame and longitudinal parity.

Nine track had both parities, plus a crc character. Data, three
frame times empty, crc frame, three frame times empty, lrc, irg.

See e.g. http://bitsavers.org/pdf/kennedy/9100/Kennedy_Model_9100_192-9100-003_75ips_Operation_and_Maintenance.pdf

De

On Sunday, February 12, 2023 at 12:11:58 AM UTC-5, Dennis Boone wrote:
> Nine track had both parities, plus a crc character. Data, three
> frame times empty, crc frame, three frame times empty, lrc, irg.
>
> See e.g. http://bitsavers.org/pdf/kennedy/9100/Kennedy_Model_9100_192-9100-003_75ips_Operation_and_Maintenance.pdf
>
> De

The appendix describes lrc and crc for NRZ1 format (and crc only for 800 CPI) but says explicitly that PE format has neither.

Clark G <clarkm.geimsler@ieeemmm.org> wrote:
>A friend of mine ran a geo-physical data processing company in Calgary. I
>visited his office around 1990 and they had the VAX 11/780 they had
>started with still running and they had started with some vector
>processing hardware attached to the 780. By then they were using high
>powered workstations for the number crunching and the 780 was just used
>for reading the raw data from the 9 track tapes the customers sent them.

The difference in floating point performance between a straight 11/780 and
a Sparcstation I was absolutely staggering. I couldn't believe how much
faster the Sparc was, and it cost less to buy than the Vax cost to maintain
for three months.

But it's true that you could hang a vector machine like the Floating Point
Systems array processors on the side of the vax and it made things much less
painful (if your code could be adapted for it).
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

On 2/12/2023 12:11 AM, Dennis Boone wrote:
> > Hmm. In 50+ years of data processing/IT/systems work, I've never heard of
> > tapes having (nonprogrammatic) CRC, that is, anything generated by the
> > hardware/firmware rather than the backup program.
>
> Seven track tape had both frame and longitudinal parity.
>
> Nine track had both parities, plus a crc character. Data, three
> frame times empty, crc frame, three frame times empty, lrc, irg.
>
> See e.g. http://bitsavers.org/pdf/kennedy/9100/Kennedy_Model_9100_192-9100-003_75ips_Operation_and_Maintenance.pdf

A quick glance indicate that this stuff is below "byte level" and down
at "electric level".

Arne

> The difference in floating point performance between a straight 11/780
> and a Sparcstation I was absolutely staggering. [...]

Well, duh. "absolutely staggering" is how I might describe the _age_
of a VAX-11/780 when the SPARCstation 1 was released (1989). A decade
later, the SPARCstation 1 might have been less impressive, too. (As I
recall, it was.)

On Sun, 2023-02-12 at 12:14 -0800, Steven Schweda wrote:
> > The difference in floating point performance between a straight
> > 11/780 and a Sparcstation I was absolutely staggering. [...]
>
>    Well, duh.  "absolutely staggering" is how I might describe the
> _age_ of a VAX-11/780 when the SPARCstation 1 was released (1989).  A
> decade later, the SPARCstation 1 might have been less impressive,
> too.  (As I recall, it was.)

I once had a SunBlade 2500 with dual 1.2GHz processors. The newer
x86_64 machines were a lot more efficient.
--
Tactical Nuclear Kittens

On 2/12/2023 3:14 PM, Steven Schweda wrote:
>> The difference in floating point performance between a straight 11/780
>> and a Sparcstation I was absolutely staggering. [...]
>
> Well, duh. "absolutely staggering" is how I might describe the _age_
> of a VAX-11/780 when the SPARCstation 1 was released (1989). A decade
> later, the SPARCstation 1 might have been less impressive, too. (As I
> recall, it was.)
>

Perhaps cost is an issue? The cost of a VAX 11/780 back then, even after 11
years, might still be on one's mind, and the reason it still might be in use.

But yeah, HW has been on a roll for a long time now, and still is doing so.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On 2/12/2023 6:08 PM, Dave Froble wrote:
> On 2/12/2023 3:14 PM, Steven Schweda wrote:
>>> The difference in floating point performance between a straight 11/780
>>> and a Sparcstation I was absolutely staggering. [...]
>>
>>    Well, duh.  "absolutely staggering" is how I might describe the _age_
>> of a VAX-11/780 when the SPARCstation 1 was released (1989).  A decade
>> later, the SPARCstation 1 might have been less impressive, too.  (As I
>> recall, it was.)
>
> Perhaps cost is an issue?  The cost of a VAX 11/780 back then, even
> after 11 years, might still be on one's mind, and the reason it still
> might be in use.

I think you are on to something.

What did a SPARCstation 1 cost in 1989?

What did a VAX 780 (let us say maxed out with memory)
cost in 1979?

The cost of the VAX 780 may have give expectations of the
system running for 10-20 years and not the 5 years
that has become custom for the "cheap" stuff.

Arne

On 2023-02-12 01:28, Arne Vajhøj wrote:
> On 2/11/2023 7:02 PM, Johnny Billquist wrote:
>> On 2023-02-11 17:26, Arne Vajhøj wrote:
>>> On 2/11/2023 11:18 AM, Tony Priborsky wrote:
>>>> I'm pretty sure it didn't write a native VMS backup saveset -
>>>> remember the CPU on the HSC was a PDP-11.   Logically the backup was
>>>> the equivalent of VMS "BACKUP/PHYSICAL".  I seem to recall that at
>>>> some point the VMS backup could restore (read) a HSC backup.
>>>
>>> So no CRC and no redundancy groups?
>>
>> Depends on how you mean that. Tapes have CRC on the data unless I
>> remember wrong. Just like disks have.
>
> As I remember 9 track tapes then it was pure data blocks
> and tape marks - no CRC.

On a software level, you don't see this anymore than you see all the
extra cruft on a disk.

> If mounted as labeled tape then some blocks were
> considered VOL and HDR blocks, but that was an OS
> feature not a drive feature.

Correct. From a hardware point of view, those are just blocks just like
anything else. No difference at all.

However, the tape have blocks. You have the tape marks, blocking and
other stuff going on, to try to ensure your data is safe and don't get
corrupted. It's all down in the hardware, but that don't mean it work
any worse/different.

Johnny

On 2/12/2023 6:44 PM, Arne Vajhøj wrote:
> On 2/12/2023 6:08 PM, Dave Froble wrote:
>> On 2/12/2023 3:14 PM, Steven Schweda wrote:
>>>> The difference in floating point performance between a straight 11/780
>>>> and a Sparcstation I was absolutely staggering. [...]
>>>
>>>    Well, duh.  "absolutely staggering" is how I might describe the _age_
>>> of a VAX-11/780 when the SPARCstation 1 was released (1989).  A decade
>>> later, the SPARCstation 1 might have been less impressive, too.  (As I
>>> recall, it was.)
>>
>> Perhaps cost is an issue?  The cost of a VAX 11/780 back then, even
>> after 11 years, might still be on one's mind, and the reason it still
>> might be in use.
>
> I think you are on to something.
>
> What did a SPARCstation 1 cost in 1989?
>
> What did a VAX 780 (let us say maxed out with memory)
> cost in 1979?
>
> The cost of the VAX 780 may have give expectations of the
> system running for 10-20 years and not the 5 years
> that has become custom for the "cheap" stuff.

Technically the 700 systems were obsolete when the
8000 systems came out mid 80's (and the 8000 systems
were obsolete when the 6000 systems came out late 80's
and early 90's and so on).

But perception and expectation for business life time
is different from technical obsolence.

Arne

On 2/12/2023 6:52 PM, Johnny Billquist wrote:
> On 2023-02-12 01:28, Arne Vajhøj wrote:
>> On 2/11/2023 7:02 PM, Johnny Billquist wrote:
>>> On 2023-02-11 17:26, Arne Vajhøj wrote:
>>>> On 2/11/2023 11:18 AM, Tony Priborsky wrote:
>>>>> I'm pretty sure it didn't write a native VMS backup saveset -
>>>>> remember the CPU on the HSC was a PDP-11.   Logically the backup was
>>>>> the equivalent of VMS "BACKUP/PHYSICAL".  I seem to recall that at
>>>>> some point the VMS backup could restore (read) a HSC backup.
>>>>
>>>> So no CRC and no redundancy groups?
>>>
>>> Depends on how you mean that. Tapes have CRC on the data unless I
>>> remember wrong. Just like disks have.
>>
>> As I remember 9 track tapes then it was pure data blocks
>> and tape marks - no CRC.
>
> On a software level, you don't see this anymore than you see all the
> extra cruft on a disk.
>
>> If mounted as labeled tape then some blocks were
>> considered VOL and HDR blocks, but that was an OS
>> feature not a drive feature.
>
> Correct. From a hardware point of view, those are just blocks just like
> anything else. No difference at all.
>
> However, the tape have blocks. You have the tape marks, blocking and
> other stuff going on, to try to ensure your data is safe and don't get
> corrupted. It's all down in the hardware, but that don't mean it work
> any worse/different.

I guess I am too far distanced from the hardware.

To me a tapes content is what DUMP will show. I have
never wondered much about how blocks and bytes was
actually stored.

Links have already been posted to NRZI, PE, GCR etc. and
I can see that there are more differences than just 800, 1600
and 6250 BPI.

Arne

=?UTF-8?Q?Arne_Vajh=c3=b8j?= <arne@vajhoej.dk> wrote:
>
>What did a SPARCstation 1 cost in 1989?

Less than a uVax 3200.
And the SLC was even cheaper.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

On 2/12/2023 6:44 PM, Arne Vajhøj wrote:
> On 2/12/2023 6:08 PM, Dave Froble wrote:
>> On 2/12/2023 3:14 PM, Steven Schweda wrote:
>>>> The difference in floating point performance between a straight 11/780
>>>> and a Sparcstation I was absolutely staggering. [...]
>>>
>>> Well, duh. "absolutely staggering" is how I might describe the _age_
>>> of a VAX-11/780 when the SPARCstation 1 was released (1989). A decade
>>> later, the SPARCstation 1 might have been less impressive, too. (As I
>>> recall, it was.)
>>
>> Perhaps cost is an issue? The cost of a VAX 11/780 back then, even after 11
>> years, might still be on one's mind, and the reason it still might be in use.
>
> I think you are on to something.
>
> What did a SPARCstation 1 cost in 1989?
>
> What did a VAX 780 (let us say maxed out with memory)
> cost in 1979?

Most likely in the neighborhood of half a million ...

Some of that will be the disk drives, which were over 20K, each ..

> The cost of the VAX 780 may have give expectations of the
> system running for 10-20 years and not the 5 years
> that has become custom for the "cheap" stuff.
>
> Arne
>
>

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486