OT: Halt and Catch Fire

[JF Mezei]

Yeah, I know , "HCF, Halt and Catch Fire" was an IBM 370 instruction in
many "hacked" university documents.

It is also an AMC TV series now available on Netflix (at least in Canada).

Seems to deal with the emergence of the PC in early 1980s.

A curious to see if Digital will ever be mentioned or ignored. So far,
one conversation had many companies listed and DEC was ignored.

[Bill Gunshannon]

In article <556e73bf$0$59523$b1db1813$2411...@news.astraweb.com>,

DEC computers didn't "Halt and Catch Fire". :-)

I seem to remember that the original HCF instruction refered to a certain
Motorolla processor that had a test mode that blocked all interupts and
put the CPU into an extremely tight internal loop which resulted in the
chip literally overheating to the point of self destruction with no way
to stop it other than yanking the plug out of the wall.

bill

--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>

[seasoned_geek]

You should actually watch it when it is on AMC. Netflix is evil.

That is actually a really good show. I really like the very Starbuck looking blonde actress. Don't think they are the same person, but never checked the credits.

While the female character makes for interesting TV, I wonder just how many women were actually "in" at that level during the early days of PCs? I mean, we had the Admiral who gave us COBOL, but I don't remember hearing about too many other Uber chip head or Uber coders back in the day.

Yes, we had quite a few women working in IT back then, but, I'm talking about working at the 'Halt and Catch Fire' level.

You know they didn't name the show after the Itanic. If they did the show would be called 'Boot and Catch Fire' or 'Fried Green Silicon'.

Roland

[Alan Frisbie]

On 06/02/2015 08:25 PM, JF Mezei wrote:
> Yeah, I know , "HCF, Halt and Catch Fire" was an IBM 370 instruction in
> many "hacked" university documents.
>

> Seems to deal with the emergence of the PC in early 1980s

I have a ditto copy (remember ditto copies?) of an instruction list from
the 1960s which had "Halt and Catch Fire", supposedly for the IBM S/360
Model 69. I am sure that it was copied from an even earlier list for the
IBM 1620, but I don't have a copy. That would have been in 1965.

Alan Frisbie

[li...@openmailbox.org]

On Wed, 03 Jun 2015 20:39:34 -0700
Alan Frisbie via Info-vax <info...@rbnsn.com> wrote:

> On 06/02/2015 08:25 PM, JF Mezei wrote:
> > Yeah, I know , "HCF, Halt and Catch Fire" was an IBM 370 instruction in
> > many "hacked" university documents.

Does that mean somebody inserted this as a joke?

> > Seems to deal with the emergence of the PC in early 1980s
>
> I have a ditto copy (remember ditto copies?) of an instruction list from
> the 1960s which had "Halt and Catch Fire", supposedly for the IBM S/360
> Model 69.

I don't think there was a Model 69 but Glen is older than I am and (usually)
remembers better.

Certainly there is no HCF instruction in the POPS for 360, 370, or later.
Not sure what this is about. I have not seen and not heard of an IBM
machine catching fire and they haven't had uptime problems since the mid
80s. Sounds like venus envy to me.

There were (are?) only 3 opcodes that began with H. If I recall two were
floating point. I don't remember what the third was.

> I am sure that it was copied from an even earlier list for the IBM 1620,
> but I don't have a copy. That would have been in 1965.

Now that very well could be. But I could not find any HCF opcode in S/360
and later doc.

--
Please DO NOT COPY ME on mailing list replies. I read the mailing list.
RSA 4096 fingerprint 7940 3F02 16D3 AFEE F2F8 ACAA 557C 4B36 98E4 4D49

[JohnF]

Alan Frisbie <Usenet0...@flying-disk.com> wrote:
> On 06/02/2015 08:25 PM, JF Mezei wrote:
>> Yeah, I know , "HCF, Halt and Catch Fire" was an IBM 370 instruction in
>> many "hacked" university documents.
>>
>> Seems to deal with the emergence of the PC in early 1980s
>
> I have a ditto copy (remember ditto copies?) of an instruction list from
> the 1960s which had "Halt and Catch Fire", supposedly for the IBM S/360
> Model 69.

As pointed out, no model 69, though I do recall a 65. And I also
recall that list posted on the front panel of either our 360/30 or 40.
Let's see, it also had RSC=read and shred card, among maybe one or two
dozen other mnemonics. But I don't have a copy.
You think you could scan your copy and post a pdf somewheres???

> I am sure that it was copied from an even earlier list for the
> IBM 1620, but I don't have a copy. That would have been in 1965.
> Alan Frisbie

Before we got a 360/30 (at NYU's School of Commerce Academic
Computing Center) we had a 1620. Maybe the list existed then,
but it definitely didn't appear at ACC until circa 1968 and
the 360/30 or 40.
--
John Forkosh ( mailto: j...@f.com where j=john and f=forkosh )

[Neil Rieck]

As others have already pointed out, there was no HCF instruction but I have been told by people who worked on those machines that you could not stop the CPU on some IBM machines: Halt really meant the CPU went into a tight NOP loop (which is were the catch-fire myth comes from).

On the flip side, there were a lot of machines which could be stopped and even "single stepped". These machines included every mini I ever worked on starting with the Interdata-70 (an IBM-360 inspired clone) and PDP-11. I remember my Interdata-70 instructor back in 1976 saying something like: "being able to stop a CPU was only possible with fully-static CPU designs". So does anyone out there have any examples of a non-static designs?

As for the AMC series, there was no Cardiff Electric. Season-1 is really about Compaq Computer Corp who had the audacity to reverse engineer the IBM-PC.

http://iq.intel.com/uncover-the-real-life-story-behind-amcs-halt-and-catch-fire/
http://www.internethistorypodcast.com/2014/05/the-incredible-true-story-behind-amcs-halt-and-catch-fire-how-compaq-cloned-ibm-and-created-an-empire/

Neil Rieck
Kitchener / Waterloo / Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/

[glen herrmannsfeldt]

li...@openmailbox.org wrote:
> On Wed, 03 Jun 2015 20:39:34 -0700
> Alan Frisbie via Info-vax <info...@rbnsn.com> wrote:

>> On 06/02/2015 08:25 PM, JF Mezei wrote:
>> > Yeah, I know , "HCF, Halt and Catch Fire" was an IBM 370
>> > instruction in many "hacked" university documents.

> Does that mean somebody inserted this as a joke?

I do remember that there were some joke instructions in those days.
Though I only remember that they were for DEC processors.

>> > Seems to deal with the emergence of the PC in early 1980s

>> I have a ditto copy (remember ditto copies?) of an instruction list from
>> the 1960s which had "Halt and Catch Fire", supposedly for the IBM S/360
>> Model 69.

> I don't think there was a Model 69 but Glen is older than
> I am and (usually) remembers better.

> Certainly there is no HCF instruction in the POPS for 360, 370, or later.
> Not sure what this is about. I have not seen and not heard of an IBM
> machine catching fire and they haven't had uptime problems since the mid
> 80s. Sounds like venus envy to me.

S/360 and successors don't have a HALT instruction. To stop the
processor, set the WAIT bit in the PSW by loading a new one with
LPSW, or from an interrupt new PSW.

> There were (are?) only 3 opcodes that began with H. If I recall two were
> floating point. I don't remember what the third was.

I don't usually think of them by their first letter. S/360 has a
HALVE instruction for floating point, HER and HDR, for short and
long floating point. There is HIO, privileged Halt I/O instruction.

So many instructions have been added in ESA/390 and now z/ that I
might expect some to start with H.

>> I am sure that it was copied from an even earlier list for the IBM 1620,
>> but I don't have a copy. That would have been in 1965.

> Now that very well could be. But I could not find any HCF opcode in S/360
> and later doc.

Now that so few people do assembly programming, it isn't as much
fun to add such instructions.

-- glen

[glen herrmannsfeldt]

Neil Rieck <n.r...@sympatico.ca> wrote:
> On Tuesday, June 2, 2015 at 11:25:53 PM UTC-4, JF Mezei wrote:
>> Yeah, I know , "HCF, Halt and Catch Fire" was an IBM 370 instruction in
>> many "hacked" university documents.

(snip)

> As others have already pointed out, there was no HCF instruction
> but I have been told by people who worked on those machines that
> you could not stop the CPU on some IBM machines: Halt really meant
> the CPU went into a tight NOP loop (which is were the catch-fire
> myth comes from).

You stop the CPU from executing instructions by setting the WAIT
bit in the PSW. 360's were leased, with the charge depending on
how much you used the system. There is a usage meter that runs when
the CPU is running, that is, when the WAIT bit is not set.

When there is nothing else to do, the CPU goes to a WAIT state, until
the next interrupt.

There are some error conditions, mostly during IPL when things
aren't going well, called disabled WAIT states. If you set WAIT
and disable all interrupts, then the system stops.

-- glen

[Chris Scheers]

IIRC, the HCF instruction was a joke in the IBM world. Along with such
things as "Execute Operator" and the even more desirable "Execute
Programmer". <grin>

The 6800 microprocessor did have an undocumented instruction that
basically turned the address bus into a 16 bit counter. This has since
been referred to as HCF.

--
-----------------------------------------------------------------------
Chris Scheers, Applied Synergy, Inc.

Voice: 817-237-3360 Internet: ch...@applied-synergy.com
Fax: 817-237-3074

[terry+go...@tmk.com]

On Thursday, June 4, 2015 at 3:27:36 AM UTC-4, JohnF wrote:
> As pointed out, no model 69, though I do recall a 65. And I also
> recall that list posted on the front panel of either our 360/30 or 40.
> Let's see, it also had RSC=read and shred card, among maybe one or two
> dozen other mnemonics. But I don't have a copy.
> You think you could scan your copy and post a pdf somewheres???

Model 69 was probably part of the joke nature of the card that listed the HCF instruction.

I never had an IBM CPU burn. I had a bunch of ITT Courier pseudo-3277 controllers that tended to smolder. I told my students "What do you expect from a company that makes both atomic bomb parts and Twinkies?". The ITT service guy was NOT amused by my putting "Thank you for not smoking" desk signs on top of each of the controllers.

RSC is only implemented by the IBM 2560 MFCM (officially "MultiFunction Card Machine", but more commonly known as the "Mother-F****** Card Mangler") due to its 2 input hoppers, 5 output stackers, and the non-deterministic path between them, where a card might randomly encounter a read station, a punch station, or an interpretation station. The initial 90-degree turn and the subsequent 180-degree turn with rotation were added simply to frustrate the user or CE who attempted to determine where the 12-car[d] pileup had happened inside the machine. https://www.google.com/search?q=ibm+2560&tbm=isch

To briefly answer some later posts...

I'm pretty sure that it was called a "hard wait" state. Machine problems during IMPL didn't get that far (since there was no WAIT instruction loaded yet). There were a row of front lamps (top right by the "Emergency Pull", IIRC) with legends like "CPU EARLY", "CPU LATE", and my favorite, "SYSTEM DAMAGE". I had a 3125 (370/125) with an add-on memory cabinet from MAI (installed with a Sawzall and hundreds of wire-wrap wires coming out of the 3125 cabinet and into the memory cabinet), where if the add-on memory was powered on during IMPL, you got both a CPU EARLY and a CPU LATE. There was only a 30-second or so period to power it on, late in IMPL, so it would be detected and sized.

Microcoded CPUs generally don't have hardware-level HALT instructions - the microcode keeps spinning to detect changes in machine state and respond accordingly. Even microcoded systems with front panels had the front panel service routine (including single step) implemented in microcode.

For the ultimate reference to things IBMish, we'd need to tempt Lynn Wheeler to post here. http://www.garlic.com/~lynn/subtopic.html

[Alan Frisbie]

On 06/04/2015 02:22 PM, terry+go...@tmk.com wrote:

> RSC is only implemented by the IBM 2560 MFCM (officially
> "MultiFunction Card Machine", but more commonly known as the
> "Mother-F****** Card Mangler") due to its 2 input hoppers,
> 5 output stackers, and the non-deterministic path between them,
> where a card might randomly encounter a read station, a punch
> station, or an interpretation station.

Our local euphemism for it was "Mother Fletcher's Card Mulcher."
I don't think I ever saw one that wasn't attached to a S/360-20.
They were a great match for each other (said in a very disparaging
tone).

[mcle...@gmail.com]

HCF and other similar instructions can be found at

http://ruthless.zathras.de/fun/top-secret/NewOpCodes.txt

or for what looks to be a longer version

https://www.physics.ohio-state.edu/~bcd/humor/instruction.set.html

Some examples ...

BW BRANCH ON WHIM
LCD Load and Clear Disk
MSGD Make Screen Go Dim
MWT MALFUNCTION WITHOUT TELLING
SHAB SHIFT A BIT
SHABM SHIFT A BIT MORE

[JF Mezei]

On 15-06-05 00:44, mcle...@gmail.com wrote:

> Some examples ...
>
> BW BRANCH ON WHIM
> LCD Load and Clear Disk
> MSGD Make Screen Go Dim
> MWT MALFUNCTION WITHOUT TELLING
> SHAB SHIFT A BIT
> SHABM SHIFT A BIT MORE

BST Backspace and stretch tape.

The version I got dates from April 1 1982, IBM Assembler G.

[JohnF]

terry+go...@tmk.com wrote:
> On Thursday, June 4, 2015 at 3:27:36 AM UTC-4, JohnF wrote:
>> As pointed out, no model 69, though I do recall a 65. And I also
>> recall that list posted on the front panel of either our 360/30 or 40.
>> Let's see, it also had RSC=read and shred card, among maybe one or two
>> dozen other mnemonics. But I don't have a copy.
>> You think you could scan your copy and post a pdf somewheres???
>
> Model 69 was probably part of the joke nature
> of the card that listed the HCF instruction.

Duh, okay, thanks, now I get it. But it's very annoying that
my computer apparently has a better sex life than I do.
I subsequently googled those instructions and found a bunch
of hits, but those lists of joke mnemonics must be later editions,
much longer than the one I recall from the late 1960's.

[terry+go...@tmk.com]

On Thursday, June 4, 2015 at 11:45:09 PM UTC-4, Alan Frisbie wrote:
> Our local euphemism for it was "Mother Fletcher's Card Mulcher."
> I don't think I ever saw one that wasn't attached to a S/360-20.
> They were a great match for each other (said in a very disparaging
> tone).

We got ours as part of an "upgrade" from a 370/115 to a /125. These are essentially the same hardware, just one of the 7 or so identical processors in the 125 (the one that handles the 370 instruction set) is clocked faster than the rest, while all are the same on the 115.

I've described our 125 as being "born under a bad sign" before - in addition to the flaky add-on memory and the MFCM, it was the only 370 we had which used a golfball (Selectric-ish, but not based on the typewriter) printing console instead of a CRT console. The continuous form paper would frequently jam in the console printer, and opening the top to clear the jam had about a 50/50 chance of crashing the system and requiring a do-over from scratch.

After that experience, all subsequent 370 systems up until we abandoned cards had a 2501 reader and 1442 punch, both of which were very reliable.

The 125 was replaced at the earliest time we could get out of the 3rd-party lease with a 138. Despite an unexpected excursion of the CPU cabinet down the street and into a cemetery where it crashed into a mausoleum and fell over during a site relocation, it performed very well and the only reason we let it go was because we needed more CPU.

The 138 was traded for a 4331. Nothing much to say about that except that I bought some extra memory cards for it, installed them, and called IBM to have them certified in the system so the system would remain on contract. This was Simply Not Done in that era and was Most Confusing to my IBM CE (who was himself an antique who had been trained on the Stretch and had a habit of smoking in the computer room and stubbing out his cigarettes in the disk drive cabinets).

A brief fling with a 4381 was followed by a 9375, which was the system that caused IBM to lose us as a customer. It never worked right, having a bunch of hardware and software problems. Example hardware problem: Flipping the "test" switch rapidly back and forth on any attached 3278 terminal would cause the processor microcode to crash and require a power cycle. Endicott spent most nights dialed into the system trying to fix it, as the proposed solution - tape signs saying "please don't rapidly flip the test switch as it will crash the system" was a non-starter in a school environment. Example software problem: The "service" (patch) tape for the VM/SP COBOL compiler wouldn't apply. IBM insisted I didn't know what I was doing and said they would send someone to install it for me, as long as we agreed to pay for it if he was successful. I agreed, with the caveat that he would remain on-site until he succeeded. IBM foolishly agreed. I arrived with my sleeping bag (I made the IBM guy sleep on the raised floor without a blanket). 3 days later he had erased the operating system by accident, gave up, and left.

I repeatedly told IBM to take it back (we had never signed the acceptance letter) but they refused. IBM had a big booth at DEXPO (remember DEXPO?) and I happened to run into a senior IBM person there and gave him the whole tale of woe, including the "27 8×10 color glossy pictures with circles and arrows and a paragraph on the back of each one". By the time DEXPO ended, I had an agreement where IBM took the system back, did not admit fault, and refunded all of our money as long as neither of us mentioned it for 5 years (long since passed).

Epilogue: I was unsuccessful in getting IBM to stop sending me service tapes (9-track) for a system I no longer had. I'd just peel the labels off and throw the tapes in my spare tape bins (giant ex-USPS hampers on wheels). One month a year or so later, IBM goofed and sent everybody service on 3480 tapes (IBM's little square tape, similar to the DEC TK70) regardless of the format listed in their customer profile. I called up the contact number on the cover letter and acted all outraged that I got the service on the wrong format, and that I needed the 9-track version RIGHT NOW. They apologized (I think they got a lot of those calls), made the tape and 4-hour messengered it to me the same day (IBM was still responsive to customers back then). The messenger showed up and as I signed for the tape, I said "you probably want to stay and watch this" as I took the tape out of the package, peeled the labels off it, and threw it into the scratch tape hamper with the other recycled tapes. We stopped receiving further service tapes after that. 8-}

[li...@openmailbox.org]

On Thu, 4 Jun 2015 11:46:42 +0000 (UTC)
glen herrmannsfeldt via Info-vax <info...@rbnsn.com> wrote:

> li...@openmailbox.org wrote:
> > There were (are?) only 3 opcodes that began with H. If I recall two were
> > floating point. I don't remember what the third was.
>
> I don't usually think of them by their first letter. S/360 has a
> HALVE instruction for floating point, HER and HDR, for short and
> long floating point. There is HIO, privileged Halt I/O instruction.

That might have been what I was thinking of but I admit to never having
used it. I figured you would remember the floating point opcode names.

> So many instructions have been added in ESA/390 and now z/ that I
> might expect some to start with H.

Yes. There have been way too many new opcodes added to z/Architecture. ESA
and S/390 were still very clean but z/Arch is a pizza with everything on
it. I don't know about the last few revisions but I have an earlyish z/OS
summary and HIO is gone but HSCH (HALT SUBCHANNEL) has been added. So,
still only three H opcodes.

[kevin....@gmail.com]

On Tuesday, June 2, 2015 at 11:25:53 PM UTC-4, JF Mezei wrote:

> A curious to see if Digital will ever be mentioned or ignored. So far,
> one conversation had many companies listed and DEC was ignored.

They briefly mention DEC and VMS. Cameron says "D-E-C-V-M-S-O-S" really fast.

[g...@gce.com]

On Tuesday, June 2, 2015 at 11:25:53 PM UTC-4, JF Mezei wrote:

There were many of these. I recall "SPW" (spacewar in one instruction) and others. The pdp1x at MIT had a long list. PDT (punch DECtape), RPM (read programmer's mind), WPM (write programmer's mind, generally returning a "writelock" error), BST (backspace and stretch tape), EPI (execute programmer immediate) and others...