In message <umms19.i7o.1@ID-201911.user.individual.net> at Fri, 29 Dec
2023 15:27:29, Frank Slootweg <this@ddress.is.invalid> writes
>Steve Hayes <hayesstw@telkomsa.net> wrote:
>> On 19 Nov 2023 19:34:29 GMT, Frank Slootweg <this@ddress.is.invalid>
>> wrote:
>>
>> >Steve Hayes <hayesstw@telkomsa.net> wrote:
>> >> On 17 Nov 2023 16:16:59 GMT, Frank Slootweg <this@ddress.is.invalid>
>> >> wrote:
>> >>
>> >> >Steve Hayes <hayesstw@telkomsa.net> wrote:
>> >> >> Someone stole my laptop computer, and I'm beginning to be concerned
>> >> >> that it may be irreplaceable.
>> >> >>
>> >> >> It was running Windows 7, 32-bit, and it seems that most, if not all,
>> >> >> laptops sold nowadays with Windows installed are 64-bit, which means
>> >> >> they won't run a lot of my software, and that means that they won't
>> >> >> allow me to access a lot of the research data I have collected over
>> >> >> the last 30 years.
>> >> >
>> >> > 64-bit Windows systems can run 32-bit software/programs just fine, so
>> >> >I think you mean you (also) have *16-bit* software/programs which you
>> >> >need to run. Correct?
>> >>
>> >> Yes, and 8-bit ones too. 32-bit Windows runs those just fine, at least
>> >> all the ones I use regularly. There are some it doesn't, but that's a
>> >> hardware rather than an O/S problem, something to do with clock speed.
>> >> Programs written in TurboPascal, for example, won't run on faster
>> >> machines.
>> >
>> > "8-bit ones" sounds a bit strange, because all (IBM-like) PCs have
>> >always been 16-bit. But perhaps you mean byte-level interpretive code or
>> >some such. Can you give some more details about these "8-bit ones"?
>>
>> I think early programs running on IBM PC DOS or MS DOS were 8-bit,
>> running on 8088 processors. The 286 and 386 ones were 16-bit.
>
> The usual confusion about the 'bit-ness'! :-)
>
> The 8088 is actually a 16-bit processor, because it has 16-bit
>registers, etc.. But the width of the *data* bus is 8-bit. The
>instruction set is named 'x86-16'.
>
><https://en.wikipedia.org/wiki/Intel_8088>

I'm pretty sure the original IBM PC used the 8086, which was a 16-bit
processor (though in some incarnations took two goes to get 16-bit data
or instructions, over an 8-bit bus). The 8086 came _after_ the 8088; not
sure what the 8088 was. The 6 in 8086 meant 16 bit, I'm pretty sure.

The 80186 was a rare beast - I don't think the core processor was much
if any more powerful (by whatever mention you like), but the chip had
some on-board bits that were normally implemented externally. The BBC
Micro "second processor" board used it, offering a weird sort of PC
(using Dr. DOS, IIRR); I don't know any other machine that used it.
>
> Likewise, the 80286 is also a 16-bit processor, but with a 16-bit data
>bus and also a 'x86-16' instruction set (with extensions).
>
><https://en.wikipedia.org/wiki/Intel_80286>
>
> The 80386 is a 32-bit processor, with a 16-bit or 32-bit databus and a
>'x86-32' instruction set.
>
><https://en.wikipedia.org/wiki/I386>
>
> As far as I know, none of the x86 processors were 8-bit processors.

For 386 and 486, the confusingly changed what "SX" and "DX" meant; on
one (I forget which), SX meant it _didn't_ have a floating-point maths
co-processor on board, DX meant it did. On the other, SX meant it had a
half-width (so 16?) bus outside the chip (so requiring two fetches to
get a word), DX had more pins on the chip. Motherboards for the one that
might or might not have the co-processor on chip often had a socket for
an external co-processor - I think that was the '387, so it's probably
the 3 series that was that. There were rumours as the time that 386SX
and 387 chips were actually 386DX chips that had failed at final test
but where the main processor or the co-processor part had passed; I've
no idea if there was any truth in that. (Certainly, a few years earlier,
the same sort of principle _had_ been used with SRAM chips - ones of a
certain size were available in two versions, one with one of the enable
pins active high and one active low, which were actually SRAMs of twice
the size which had failed final test but either the upper or lower half
worked.)
>
> So consequently, also the IBM PC DOS and MS-DOS programs were 16-bit.
>
>[No longer relevant stuff deleted.]
>
>> Anyway, I have now acquired a 2nd-hand Dell with a 30-bit Windows 10
>> OS, and my DOS programs appear to run on it, so I'm not going to need
>> VirtualBox just yet.
>
> Great! Good outcome!

Indeed! (32 bit I presume!)
>
>> But thanks to everyone who replied on the Windows question (the time
>> discussion was less useful).
>
> You're welcome.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

he was eventually struck off by the BMA in 1968 for not knowing his gluteus
maximus from his humerus.

J. P. Gilliver <G6JPG@255soft.uk> wrote:
> In message <umms19.i7o.1@ID-201911.user.individual.net> at Fri, 29 Dec
> 2023 15:27:29, Frank Slootweg <this@ddress.is.invalid> writes
> >Steve Hayes <hayesstw@telkomsa.net> wrote:
> >> On 19 Nov 2023 19:34:29 GMT, Frank Slootweg <this@ddress.is.invalid>
> >> wrote:
> >>
> >> >Steve Hayes <hayesstw@telkomsa.net> wrote:
> >> >> On 17 Nov 2023 16:16:59 GMT, Frank Slootweg <this@ddress.is.invalid>
> >> >> wrote:
> >> >>
> >> >> >Steve Hayes <hayesstw@telkomsa.net> wrote:
> >> >> >> Someone stole my laptop computer, and I'm beginning to be concerned
> >> >> >> that it may be irreplaceable.
> >> >> >>
> >> >> >> It was running Windows 7, 32-bit, and it seems that most, if not all,
> >> >> >> laptops sold nowadays with Windows installed are 64-bit, which means
> >> >> >> they won't run a lot of my software, and that means that they won't
> >> >> >> allow me to access a lot of the research data I have collected over
> >> >> >> the last 30 years.
> >> >> >
> >> >> > 64-bit Windows systems can run 32-bit software/programs just fine, so
> >> >> >I think you mean you (also) have *16-bit* software/programs which you
> >> >> >need to run. Correct?
> >> >>
> >> >> Yes, and 8-bit ones too. 32-bit Windows runs those just fine, at least
> >> >> all the ones I use regularly. There are some it doesn't, but that's a
> >> >> hardware rather than an O/S problem, something to do with clock speed.
> >> >> Programs written in TurboPascal, for example, won't run on faster
> >> >> machines.
> >> >
> >> > "8-bit ones" sounds a bit strange, because all (IBM-like) PCs have
> >> >always been 16-bit. But perhaps you mean byte-level interpretive code or
> >> >some such. Can you give some more details about these "8-bit ones"?
> >>
> >> I think early programs running on IBM PC DOS or MS DOS were 8-bit,
> >> running on 8088 processors. The 286 and 386 ones were 16-bit.
> >
> > The usual confusion about the 'bit-ness'! :-)
> >
> > The 8088 is actually a 16-bit processor, because it has 16-bit
> >registers, etc.. But the width of the *data* bus is 8-bit. The
> >instruction set is named 'x86-16'.
> >
> ><https://en.wikipedia.org/wiki/Intel_8088>
>
> I'm pretty sure the original IBM PC used the 8086, which was a 16-bit
> processor (though in some incarnations took two goes to get 16-bit data
> or instructions, over an 8-bit bus).

I commented on the 8088, because that's what Steve mentioned, not
specifically in the context of the IBM PC. But read on! :-)

> The 8086 came _after_ the 8088; not
> sure what the 8088 was. The 6 in 8086 meant 16 bit, I'm pretty sure.

Nope, as to be expected, the 8088 came after the 8086. For a quick
explanation of the differences and timeline, hover over '8086' on the
(start of the) second line of the first paragraph on
<https://en.wikipedia.org/wiki/Intel_8088>

BTW, this popup and the 'Selection for use in the IBM PC' says that
the original IBM PC used the 8088, not the 8086 as you thought.

> The 80186 was a rare beast - I don't think the core processor was much
> if any more powerful (by whatever mention you like), but the chip had
> some on-board bits that were normally implemented externally. The BBC
> Micro "second processor" board used it, offering a weird sort of PC
> (using Dr. DOS, IIRR); I don't know any other machine that used it.
> >
> > Likewise, the 80286 is also a 16-bit processor, but with a 16-bit data
> >bus and also a 'x86-16' instruction set (with extensions).
> >
> ><https://en.wikipedia.org/wiki/Intel_80286>
> >
> > The 80386 is a 32-bit processor, with a 16-bit or 32-bit databus and a
> >'x86-32' instruction set.
> >
> ><https://en.wikipedia.org/wiki/I386>
> >
> > As far as I know, none of the x86 processors were 8-bit processors.
>
> For 386 and 486, the confusingly changed what "SX" and "DX" meant; on
> one (I forget which), SX meant it _didn't_ have a floating-point maths
> co-processor on board, DX meant it did. On the other, SX meant it had a
> half-width (so 16?) bus outside the chip (so requiring two fetches to
> get a word), DX had more pins on the chip. Motherboards for the one that
> might or might not have the co-processor on chip often had a socket for
> an external co-processor - I think that was the '387, so it's probably
> the 3 series that was that. There were rumours as the time that 386SX
> and 387 chips were actually 386DX chips that had failed at final test
> but where the main processor or the co-processor part had passed; I've
> no idea if there was any truth in that. (Certainly, a few years earlier,
> the same sort of principle _had_ been used with SRAM chips - ones of a
> certain size were available in two versions, one with one of the enable
> pins active high and one active low, which were actually SRAMs of twice
> the size which had failed final test but either the upper or lower half
> worked.)

If you or anyone want(s) to know which is which, just type the numbers
into the Wikipedia search box and all will be revealed! :-)

BTW, my first PC had a (16 MHz) 386DX, which had a 32-bit data bus
(instead of 16-bit for the 386SX).

<https://en.wikipedia.org/wiki/I386>

> > So consequently, also the IBM PC DOS and MS-DOS programs were 16-bit.
> >
> >[No longer relevant stuff deleted.]
> >
> >> Anyway, I have now acquired a 2nd-hand Dell with a 30-bit Windows 10
> >> OS, and my DOS programs appear to run on it, so I'm not going to need
> >> VirtualBox just yet.
> >
> > Great! Good outcome!
>
> Indeed! (32 bit I presume!)

Nah, it's one of them newfangled 30-bit ones! Runs all software ever
written and to be written and cleans the kitchen sink while doing it!

> >
> >> But thanks to everyone who replied on the Windows question (the time
> >> discussion was less useful).
> >
> > You're welcome.

On 12/29/2023 1:26 PM, Mark Lloyd wrote:
> [snip]
>
>> They often have a USB slot in them too, so you aren't "wasting" a USB header.
>
> The headers usually provide 2 ports. The card reader just needs 1 so its not too hard to extend the other one to the outside of the case.
>

You can sometimes purchase equipment that has a 1x5 connector.
But the odds are slim, of purchasing a second device that also has a 1x5
to sit next to it on a header.

You can see the header rows, aren't exactly the same. If an equipment
has a plastic facade on the front, it *might* be using Shield Ground for its Shield Ground :-)

https://web.archive.org/web/20070108162027im_/http://www.frontx.com/cpx108_2p3.gif

If your equipment has a 2x5 on the end, then the details are taken care
of for you, whatever electrical connection they want to use for grounding.

Grounding to chassis, is done best if done near the faceplate.
(Metal on your new toy, touches metal on the chassis.)
Otherwise, an ESD discharge goes down the S-Ground wire and the
field couples into the other wires in the cable assembly,
and could cause an upset.

This is why, if someone has the choice of working on the front
of the computer, or on the back, I tell them to use the back, because
the shield grounding scheme on the back is normally a better one.
Some of the setups on the front of the computer, no electrical engineer
helped the fools design it. Some computer case manufacturers have only
"metal bashers" on staff, and they've made some pretty bad mistakes
on port wiring and setup on case fronts. Antec has had a few incidents (mis-wired
interfaces, strangely none of the errors made, guaranteed destruction :-) ).
This is one of the reasons, if I received an Antec case years ago,
I ripped the front panel wiring out, as the "first step". That's because
I did not want to use a multimeter to check the wiring. I have better
things to do.

Paul

In message <umo3cd$15rus$1@dont-email.me> at Fri, 29 Dec 2023 22:38:53,
Paul <nospam@needed.invalid> writes
>On 12/29/2023 1:26 PM, Mark Lloyd wrote:
>> [snip]
>>
>>> They often have a USB slot in them too, so you aren't "wasting" a
>>>USB header.
>>
>> The headers usually provide 2 ports. The card reader just needs 1 so
>>its not too hard to extend the other one to the outside of the case.
>>
>
>You can sometimes purchase equipment that has a 1x5 connector.
>But the odds are slim, of purchasing a second device that also has a 1x5
>to sit next to it on a header.
[]
Sometimes you get a 2×5 connector which you can _see_ is only wired
along one side; the provider chose to use that to ensure you plug it in
the right way round. You can then - if there is room adjacent to the
header on the board - plug it in offset by one, leaving the other port
still usable (though not by another offset plug if the header is wired
as per the header Paul has linked to, unless it happens to have used the
other row).
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

I don't like that word [atheist]; it implies that there's a god not to believe
in - Eric Idle, quoted in RT 2016/12/10-16

"J. P. Gilliver" <G6JPG@255soft.uk> wrote:

>For 386 and 486, the confusingly changed what "SX" and "DX" meant; on
>one (I forget which), SX meant it _didn't_ have a floating-point maths
>co-processor on board, DX meant it did. On the other, SX meant it had a
>half-width (so 16?) bus outside the chip (so requiring two fetches to

The 486 was the first Intel chip to have the numeric coprocessor
onboard. Intel wanted to prese4ve the "SX" price point, so they
produced a 486SX chip which was identical to the DX except that the
numeric coprocessor was disabled! Machines sold with this chip had an
empty socket where you could plug in a 486DX chip to get a coproc. So
once you did that, you could unplug the SX chip and use it elsewhere,
right? WRONG!!! It was set up so that the DX in those machines
wouldn't work unless the SX was plugged in, doing nothing.

--
Tim Slattery
timslattery <at> utexas <dot> edu

In message <3fj0pi1he28cfgbul4lbmu4jua526btugq@4ax.com> at Sat, 30 Dec
2023 12:10:57, Tim Slattery <TimSlattery@utexas.edu> writes
>"J. P. Gilliver" <G6JPG@255soft.uk> wrote:
>
>
>>For 386 and 486, the confusingly changed what "SX" and "DX" meant; on
>>one (I forget which), SX meant it _didn't_ have a floating-point maths
>>co-processor on board, DX meant it did. On the other, SX meant it had a
>>half-width (so 16?) bus outside the chip (so requiring two fetches to
>
>The 486 was the first Intel chip to have the numeric coprocessor
>onboard. Intel wanted to prese4ve the "SX" price point, so they
>produced a 486SX chip which was identical to the DX except that the
>numeric coprocessor was disabled! Machines sold with this chip had an
>empty socket where you could plug in a 486DX chip to get a coproc. So
>once you did that, you could unplug the SX chip and use it elsewhere,
>right? WRONG!!! It was set up so that the DX in those machines
>wouldn't work unless the SX was plugged in, doing nothing.
>
Did anyone ever manage to "crack" the 486SX or the 487 to enable the
disabled part, or make it work without the other?
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Q. How much is 2 + 2?
A. Thank you so much for asking your question.
Are you still having this problem? I'll be delighted to help you. Please
restate the problem twice and include your Windows version along with
all error logs.
- Mayayana in alt.windows7.general, 2018-11-1

On 12/30/23 02:30, J. P. Gilliver wrote:

[snip]

> Sometimes you get a 2×5 connector which you can _see_ is only wired
> along one side; the provider chose to use that to ensure you plug it in
> the right way round. You can then - if there is room adjacent to the
> header on the board - plug it in offset by one, leaving the other port
> still usable (though not by another offset plug if the header is wired
> as per the header Paul has linked to, unless it happens to have used the
> other row).

I have seen one device like that (uses only one port, but still has a
2x5 connector). This was a CPU cooler with LEDs on it that an be
controlled through USB.

--
Mark Lloyd
http://notstupid.us/

"Our royal flush beats your full house." - Roto-Rooter

On 12/30/2023 3:27 PM, J. P. Gilliver wrote:
> In message <3fj0pi1he28cfgbul4lbmu4jua526btugq@4ax.com> at Sat, 30 Dec 2023 12:10:57, Tim Slattery <TimSlattery@utexas.edu> writes
>> "J. P. Gilliver" <G6JPG@255soft.uk> wrote:
>>
>>
>>> For 386 and 486, the confusingly changed what "SX" and "DX" meant; on
>>> one (I forget which), SX meant it _didn't_ have a floating-point maths
>>> co-processor on board, DX meant it did. On the other, SX meant it had a
>>> half-width (so 16?) bus outside the chip (so requiring two fetches to
>>
>> The 486 was the first Intel chip to have the numeric coprocessor
>> onboard. Intel wanted to prese4ve the "SX" price point, so they
>> produced a 486SX chip which was identical to the DX except that the
>> numeric coprocessor was disabled! Machines sold with this chip had an
>> empty socket where you could plug in a 486DX chip to get a coproc. So
>> once you did that, you could unplug the SX chip and use it elsewhere,
>> right? WRONG!!! It was set up so that the DX in those machines
>> wouldn't work unless the SX was plugged in, doing nothing.
>>
> Did anyone ever manage to "crack" the 486SX or the 487 to enable the disabled part, or make it work without the other?

You will remember that these were simpler times.

Intel has much better mechanisms "to enforce this or that" today.

https://www.vogons.org/viewtopic.php?t=68471

Today we have sharks with lasers on their heads. Back then,
all we had was sharks.

Paul

In message <umqps0$1kfau$1@dont-email.me> at Sat, 30 Dec 2023 23:14:56,
Paul <nospam@needed.invalid> writes
>On 12/30/2023 3:27 PM, J. P. Gilliver wrote:
>> In message <3fj0pi1he28cfgbul4lbmu4jua526btugq@4ax.com> at Sat, 30
>>Dec 2023 12:10:57, Tim Slattery <TimSlattery@utexas.edu> writes
[]
>>> The 486 was the first Intel chip to have the numeric coprocessor
>>> onboard. Intel wanted to prese4ve the "SX" price point, so they
>>> produced a 486SX chip which was identical to the DX except that the
>>> numeric coprocessor was disabled! Machines sold with this chip had an
>>> empty socket where you could plug in a 486DX chip to get a coproc. So
>>> once you did that, you could unplug the SX chip and use it elsewhere,
>>> right? WRONG!!! It was set up so that the DX in those machines
>>> wouldn't work unless the SX was plugged in, doing nothing.
>>>
>> Did anyone ever manage to "crack" the 486SX or the 487 to enable the
>>disabled part, or make it work without the other?
>
>You will remember that these were simpler times.
>
>Intel has much better mechanisms "to enforce this or that" today.
>
> https://www.vogons.org/viewtopic.php?t=68471
>
>Today we have sharks with lasers on their heads. Back then,
>all we had was sharks.
>
> Paul

Reading that reminds me that about then was the start of different clock
speeds internally and externally - the DX2 variant.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

The way I see it, if you want the rainbow, you gotta put up with the rain
(Dolly Parton)

J. P. Gilliver wrote on 31/12/23 7:27 am:
> In message <3fj0pi1he28cfgbul4lbmu4jua526btugq@4ax.com> at Sat, 30
> Dec 2023 12:10:57, Tim Slattery <TimSlattery@utexas.edu> writes
>> "J. P. Gilliver" <G6JPG@255soft.uk> wrote:
>>
>>> For 386 and 486, the confusingly changed what "SX" and "DX"
>>> meant; on one (I forget which), SX meant it _didn't_ have a
>>> floating-point maths co-processor on board, DX meant it did. On
>>> the other, SX meant it had a half-width (so 16?) bus outside the
>>> chip (so requiring two fetches to
>>
>> The 486 was the first Intel chip to have the numeric coprocessor
>> onboard. Intel wanted to prese4ve the "SX" price point, so they
>> produced a 486SX chip which was identical to the DX except that
>> the numeric coprocessor was disabled! Machines sold with this chip
>> had an empty socket where you could plug in a 486DX chip to get a
>> coproc. So once you did that, you could unplug the SX chip and use
>> it elsewhere, right? WRONG!!! It was set up so that the DX in those
>> machines wouldn't work unless the SX was plugged in, doing
>> nothing.
>>
> Did anyone ever manage to "crack" the 486SX or the 487 to enable the
> disabled part, or make it work without the other?

"487"?? All DuckDuckGo shows seems to concern a Californian Penal Code
clause 487!! ;-P
--
Daniel

In message <umrm4d$1nhbb$2@dont-email.me> at Sun, 31 Dec 2023 23:17:19,
Daniel65 <daniel47@nomail.afraid.org> writes
>J. P. Gilliver wrote on 31/12/23 7:27 am:
>> In message <3fj0pi1he28cfgbul4lbmu4jua526btugq@4ax.com> at Sat, 30
>> Dec 2023 12:10:57, Tim Slattery <TimSlattery@utexas.edu> writes
>>> "J. P. Gilliver" <G6JPG@255soft.uk> wrote:
>>>
>>>> For 386 and 486, the confusingly changed what "SX" and "DX"
>>>> meant; on one (I forget which), SX meant it _didn't_ have a
>>>> floating-point maths co-processor on board, DX meant it did. On
>>>> the other, SX meant it had a half-width (so 16?) bus outside the
>>>> chip (so requiring two fetches to
>>> The 486 was the first Intel chip to have the numeric coprocessor
>>>onboard. Intel wanted to prese4ve the "SX" price point, so they
>>>produced a 486SX chip which was identical to the DX except that
>>> the numeric coprocessor was disabled! Machines sold with this chip
>>> had an empty socket where you could plug in a 486DX chip to get a
>>> coproc. So once you did that, you could unplug the SX chip and use
>>> it elsewhere, right? WRONG!!! It was set up so that the DX in those
>>> machines wouldn't work unless the SX was plugged in, doing
>>> nothing.
>>>
>> Did anyone ever manage to "crack" the 486SX or the 487 to enable the
>> disabled part, or make it work without the other?
>
>"487"?? All DuckDuckGo shows seems to concern a Californian Penal Code
>clause 487!! ;-P

Maybe it was 486DX as Tim says. I had _thought_ the '387 was the
co-processor for that series.

I'm not sure when they started to drop the "80" from (e. g.) 80386. I
know they started using names around the time of the '586, alias
Pentium, because someone in charge of the administration of trademarks
said, basically, no more trademarking just numbers. (It wasn't just
Intel - other manufacturers had to invent names too; I remember one chip
called "roboclock"; I think it was Maxim or IDT.) Presumably that's why
things moved to Pentium II, etc., rather than 686.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

.... the pleasure of the mind is an amazing thing. My life has been driven by
the satisfaction of curiosity. - Jeremy Paxman (being interviewed by Anne
Widdecombe), Radio Times, 2-8 July 2011.

On 12/31/2023 7:37 AM, J. P. Gilliver wrote:

> Maybe it was 486DX as Tim says. I had _thought_ the '387 was the co-processor for that series.
>
> I'm not sure when they started to drop the "80" from (e. g.) 80386. I know they started using names around the time of the '586, alias Pentium, because someone in charge of the administration of trademarks said, basically, no more trademarking just numbers. (It wasn't just Intel - other manufacturers had to invent names too; I remember one chip called "roboclock"; I think it was Maxim or IDT.) Presumably that's why things moved to Pentium II, etc., rather than 686.

It's possible for a company to continue using its old
numbering scheme, for labeling at chip level.

Here, the lucky 8 is still in usage, for a part of it. Southbridges.
These chips did not typically need heatsinks, so you could spot
a part number like this on your Pentium 4 motherboard.

https://en.wikipedia.org/wiki/I%2FO_Controller_Hub

82801ER (ICH5R) RAID

*******

Someone has saved the musty old books. Good on 'em.
That means these are likely scanned by hand. The topical
ones for this discussion, would have to be carefully picked,
to get a good cross-section of parts in use. At one time,
things like the timer chip, were a separate chip. Whereas
today, an emulation of the chip is inside a Southbridge or
a PCH. Just as the Southbridge has a gate level emulation
of a Motorala clock chip (the RTC time section).

http://www.bitsavers.org/components/intel/_dataBooks/

I used to collect those, early on, but later it did not
align with what I was doing.

Once the industry became more adept at using PDF, the
fascination with catalogs disappeared. We actually had
several very large rooms, with all those databooks collected
in them. But that library was too far from me, to be
of any practical use.

I still have a databook from 1972 here. The sentimental
aspect was, as a kid, I wrote a letter to Fairchild, asking
them if I could have a databook. And they actually sent
me a databook! Surprised the hell out of me. I think it
cost them $4.00 in postage to send that. That was the 7400
series of TTL or so. I still have a few bins of the critters.
A number of the companies from that era, did not survive.
Texas Instruments is still around.

Paul

In message <ums57t$1pjk8$1@dont-email.me> at Sun, 31 Dec 2023 11:35:08,
Paul <nospam@needed.invalid> writes
>On 12/31/2023 7:37 AM, J. P. Gilliver wrote:
>
>> Maybe it was 486DX as Tim says. I had _thought_ the '387 was the
>>co-processor for that series.
>>
>> I'm not sure when they started to drop the "80" from (e. g.) 80386. I
>>know they started using names around the time of the '586, alias
>>Pentium, because someone in charge of the administration of trademarks
>>said, basically, no more trademarking just numbers. (It wasn't just
>>Intel - other manufacturers had to invent names too; I remember one
>>chip called "roboclock"; I think it was Maxim or IDT.) Presumably
>>that's why things moved to Pentium II, etc., rather than 686.
>
>It's possible for a company to continue using its old
>numbering scheme, for labeling at chip level.
>
>Here, the lucky 8 is still in usage, for a part of it. Southbridges.
>These chips did not typically need heatsinks, so you could spot
>a part number like this on your Pentium 4 motherboard.
>
>https://en.wikipedia.org/wiki/I%2FO_Controller_Hub

Wow, that's an amazing resource!
>
> 82801ER (ICH5R) RAID
>
>*******
>
>Someone has saved the musty old books. Good on 'em.
[]
>I used to collect those, early on, but later it did not
>align with what I was doing.
>
>Once the industry became more adept at using PDF, the
>fascination with catalogs disappeared. We actually had
>several very large rooms, with all those databooks collected
>in them. But that library was too far from me, to be
>of any practical use.
>
>I still have a databook from 1972 here. The sentimental
>aspect was, as a kid, I wrote a letter to Fairchild, asking
>them if I could have a databook. And they actually sent
>me a databook! Surprised the hell out of me. I think it
>cost them $4.00 in postage to send that. That was the 7400
>series of TTL or so. I still have a few bins of the critters.

I used to collect the TTL books - had quite a large collection of them,
and knowledge of the numbering too. I only got round to throwing them
out about a year ago - I kept the earliest, as I'd actually bought it
(most of the others I'd got from work, either "acquired" or rescued when
being thrown out), and _one or two_ others (and wallcharts).

>A number of the companies from that era, did not survive.
>Texas Instruments is still around.

Initially I associate it with National (Semiconductor) - blue books, and
Texas Instruments (yellow or orange). Mullard the 11000 (centre power)
series, IDT the clever ones and also low-power-but-still-fast ACT
competitors to CMOS 4000 (which use a _lot_ less power and could be run
on a wide range of voltages, but weren't much for speed). Ah, happy
times ... oh, and - I forget who did them - somebody did single gates,
so you didn't have to have the four (etc.) you got in a 7400.
>
> Paul
John
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Can a blue man sing the whites?

[snip]

>> Did anyone ever manage to "crack" the 486SX or the 487 to enable the
>>  disabled part, or make it work without the other?
>
> "487"?? All DuckDuckGo shows seems to concern a Californian Penal Code
> clause 487!! ;-P

80487. The add-on FPU for the 80486sx, that was actually a 486dx.

--
Mark Lloyd
http://notstupid.us/

"I wake up every morning and I wish I were dead, and so does Jim."
[Tammy Fae Bakker]

On 12/31/23 13:15, J. P. Gilliver wrote:

[snip]

> I used to collect the TTL books - had quite a large collection of them,
> and knowledge of the numbering too. I only got round to throwing them
> out about a year ago - I kept the earliest, as I'd actually bought it
> (most of the others I'd got from work, either "acquired" or rescued when
> being thrown out), and _one or two_ others (and wallcharts).

I have an Intel CPU data book that is older than the 80486. One CPU I
hadn't heard of before was the 80376. It's a version of the 80386
without real mode.

--
Mark Lloyd
http://notstupid.us/

"I wake up every morning and I wish I were dead, and so does Jim."
[Tammy Fae Bakker]

On 12/31/2023 3:39 PM, Mark Lloyd wrote:
>
> [snip]
>
>>> Did anyone ever manage to "crack" the 486SX or the 487 to enable the
>>>  disabled part, or make it work without the other?
>>
>> "487"?? All DuckDuckGo shows seems to concern a Californian Penal Code
>> clause 487!! ;-P
>
> 80487. The add-on FPU for the 80486sx, that was actually a 486dx.
>

These may have worked via F line.

"$Fxxx, F-Line instructions, emulating co-pro on the systems w/o FPU,
or propagating directly to the real co-pro on systems with FPU in the socket."

The accelerator, may have been watching the bus as the main processor
accessed stuff. And if an instruction with an F in the appropriate place
showed up, the FPU knew it had a job to do. That was part of the coordination.
But I never worked on anything like that, and that info likely came
from one of my magazines at the time.

For a thing like that to work, the CPU could only have the one core.
Back in those days, instruction traces were a wee bit easier to arrange,
than they are on modern CPU sockets.

Paul

J. P. Gilliver wrote on 31/12/23 11:37 pm:
> In message <umrm4d$1nhbb$2@dont-email.me> at Sun, 31 Dec 2023
> 23:17:19, Daniel65 <daniel47@nomail.afraid.org> writes
>> J. P. Gilliver wrote on 31/12/23 7:27 am:
>>> In message <3fj0pi1he28cfgbul4lbmu4jua526btugq@4ax.com> at Sat,
>>> 30 Dec 2023 12:10:57, Tim Slattery <TimSlattery@utexas.edu>
>>> writes
>>>> "J. P. Gilliver" <G6JPG@255soft.uk> wrote:
>>>>
>>>>> For 386 and 486, the confusingly changed what "SX" and "DX"
>>>>> meant; on one (I forget which), SX meant it _didn't_ have a
>>>>> floating-point maths co-processor on board, DX meant it did.
>>>>> On the other, SX meant it had a half-width (so 16?) bus
>>>>> outside the chip (so requiring two fetches to
>>>> The 486 was the first Intel chip to have the numeric
>>>> coprocessor onboard. Intel wanted to prese4ve the "SX" price
>>>> point, so they produced a 486SX chip which was identical to the
>>>> DX except that the numeric coprocessor was disabled! Machines
>>>> sold with this chip had an empty socket where you could plug in
>>>> a 486DX chip to get a coproc. So once you did that, you could
>>>> unplug the SX chip and use it elsewhere, right? WRONG!!! It was
>>>> set up so that the DX in those machines wouldn't work unless
>>>> the SX was plugged in, doing nothing.
>>>>
>>> Did anyone ever manage to "crack" the 486SX or the 487 to enable
>>> the disabled part, or make it work without the other?
>>
>> "487"?? All DuckDuckGo shows seems to concern a Californian Penal
>> Code clause 487!! ;-P
>
> Maybe it was 486DX as Tim says. I had _thought_ the '387 was the
> co-processor for that series.
>
> I'm not sure when they started to drop the "80" from (e. g.) 80386. I
> know they started using names around the time of the '586, alias
> Pentium, because someone in charge of the administration of
> trademarks said, basically, no more trademarking just numbers. (It
> wasn't just Intel - other manufacturers had to invent names too; I
> remember one chip called "roboclock"; I think it was Maxim or IDT.)
> Presumably that's why things moved to Pentium II, etc., rather than
> 686.

(Rumour-mung Rumour-mung) The story I was told, way-back-when, was that
in an effort to wipe out Apple and Commodore, etc, Intel virtually gave
anybody the Rites and the Chip designs to produce the 80186, 80286,
80386 and 80486 Chips.

Then, having effectively achieved their Aim, when Intel produced the
80586, a.k.a. the Pentium chip, IBM required Royalties-type stuff (shut
the gate), so some manufactures (AMD, etc.) continued development
independently, to produce the 586 and 686 chips.

It may have also been the case that, as seperate companies were doing
their own development, some 586/686 chips had different pin-outs and/or
Op-Codes to other 586/686 chips!!

Something like that.
--
Daniel

Mark Lloyd <not.email@all.invalid> wrote:
> On 12/31/23 13:15, J. P. Gilliver wrote:
>
> [snip]
>
> > I used to collect the TTL books - had quite a large collection of them,
> > and knowledge of the numbering too. I only got round to throwing them
> > out about a year ago - I kept the earliest, as I'd actually bought it
> > (most of the others I'd got from work, either "acquired" or rescued when
> > being thrown out), and _one or two_ others (and wallcharts).
>
> I have an Intel CPU data book that is older than the 80486. One CPU I
> hadn't heard of before was the 80376. It's a version of the 80386
> without real mode.

Correct:

<https://en.wikipedia.org/wiki/Intel_80376>

"The Intel 80376, introduced January 16, 1989, was a variant of the
Intel 80386SX intended for embedded systems. It differed from the 80386
in not supporting real mode (the processor booted directly into 32-bit
protected mode)[1] and having no support for paging in the MMU. The 376
was available at 16 or 20 MHz clock rates."

On Fri, 29 Dec 2023 11:31:31 -0500
Tim Slattery <TimSlattery@utexas.edu> wrote:

> Steve Hayes <hayesstw@telkomsa.net> wrote:
>
>
> >I think early programs running on IBM PC DOS or MS DOS were 8-bit,
> >running on 8088 processors. The 286 and 386 ones were 16-bit.
>
> Not so. The original IBM-PCs were 16-bit machines. They used a kludge
> to implement a 20-bit address space, allowing access to one megabyte
> of RAM. You may remember that 340KB of that was reserved for the
> operating system, leaving 640KB for user program.

No to that 2nd sentence; the upper address space was used for direct
addressing the screen buffer (0xB000 for mono, 0xB800 for text/CGA, 0xA000+
for higher modes)), extension card interfaces, and ROM (E000 & F000) - I
suppose you could argue that a ROM BASIC is an operating system, but I
wouldn't accept that. Certainly user programs were only allowed a max
memory footprint of 640k - in practice much less due to DOS overhead.

> 1970's vintage machines, such as Cromemco, Zylog, etc, etc, were 8
> bits. I'm a bit foggy on their addressing schemes, but at least some
> of them could switch between banks of 64KB each.
>
> The 80286 was basically 16-bits, but implemented "protected mode"
> which allowed access to 16MB. Windows programers (if they're old
> enough) may remember using "GlobalAlloc" and "GlobalFree" calls. Those
> manipulated the Global Allocation Table, a 80286 protected mode
> hardware kludge that kept track of all that RAM.
>
> The 80386 was Intel's first true 32-bit machine. Windows 3.0 386
> version ran in 16-bit 80286 protected mode though. It took a while for
> Windows to catch up with 32-bit processors.
>
> --
> Tim Slattery
> timslattery <at> utexas <dot> edu

--
Bah, and indeed Humbug.

On Fri, 29 Dec 2023 04:21:24 -0500, Paul <nospam@needed.invalid>
wrote:

>You can get SD card readers, in the form of USB sticks with a "hole in the side".
>That's what I use for my camera. Mine will only read media up to 32GB in size.
>
> https://c1.neweggimages.com/ProductImageCompressAll1280/20-208-939-05.jpg

Thanks very much, I'll look for such a thing.

--
Steve Hayes from Tshwane, South Africa
Web: http://www.khanya.org.za/stevesig.htm
Blog: http://khanya.wordpress.com
E-mail - see web page, or parse: shayes at dunelm full stop org full stop uk