Has anyone here got access to a ridiculous computer? Something like:

CHAR_BIT == 9
PADDING_BITS(int) != 0
NUMBER_SYSTEM == SIGN_MAGNITUDE
Null pointer bit pattern == All ones

I'm writing some fully-portable code and would love to actually compile it
for, and test it on, a very strange machine :-D

Or does anyone know of a particular machine I could get my hands on very
cheaply just for playing around with, something very strange which I can
get a compliant C89 compiler for? An old supercomputer perhaps? The main
thing I'd want is a strange value for CHAR_BIT, but I'd be ecstatic if the
integer types had padding aswell!

--
Tomás Ó hÉilidhe

"Tomás Ó hÉilidhe" <toe--at--lavabit.com> writes:
> Has anyone here got access to a ridiculous computer? Something like:
>
> CHAR_BIT == 9
> PADDING_BITS(int) != 0
> NUMBER_SYSTEM == SIGN_MAGNITUDE
> Null pointer bit pattern == All ones
>
> I'm writing some fully-portable code and would love to actually compile it
> for, and test it on, a very strange machine :-D
>
> Or does anyone know of a particular machine I could get my hands on very
> cheaply just for playing around with, something very strange which I can
> get a compliant C89 compiler for? An old supercomputer perhaps? The main
> thing I'd want is a strange value for CHAR_BIT, but I'd be ecstatic if the
> integer types had padding aswell!

Your best bet for CHAR_BIT > 8 is probably a DSP (Digital Signal
Processor), but I think those systems tend not to have hosted C
implementations (i.e., much of the standard library may be missing).

You *might* be able to obtain an old working supercomputer for not too
much money, but the power and air conditioning costs are likely to be
huge. In any case, on the relatively old supercomputers I've used
(Cray T90), CHAR_BIT==8 (the compiler went to considerable effort to
make this work, even though the hardware didn't support 8-bit memory
access. But I believe at least some of the predefined integer types
did have padding bits.

cray-cyber.org offers free access to some old supercomputers; some of
them are available 24/7, others are not.

--
Keith Thompson (The_Other_Keith) <kst-u--at--mib.org>
[...]
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

In article <Xns9A18EBF80CB1Etoelavabitcom--at--194.125.133.14>,
Tomás Ó hÉilidhe <toe--at--lavabit.com> wrote:

>Has anyone here got access to a ridiculous computer? Something like:

> CHAR_BIT == 9

Possibly you could use one of the emulators at simh.trailing-edge.com.

"SIMH implements simulators for:

* Data General Nova, Eclipse
* Digital Equipment Corporation PDP-1, PDP-4, PDP-7, PDP-8, PDP-9, PDP-10, PDP-11, PDP-15, VAX
* GRI Corporation GRI-909
* IBM 1401, 1620, 1130, 7090/7094, System 3
* Interdata (Perkin-Elmer) 16b and 32b systems
* Hewlett-Packard 2114, 2115, 2116, 2100, 21MX
* Honeywell H316/H516
* MITS Altair 8800, with both 8080 and Z80
* Royal-Mcbee LGP-30, LGP-21
* Scientific Data Systems SDS 940"
--
"There are some ideas so wrong that only a very intelligent person
could believe in them." -- George Orwell

"Tomï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï ¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ " wrote:
>
> Has anyone here got access to a ridiculous computer? Something like:
>
> CHAR_BIT == 9
> PADDING_BITS(int) != 0
> NUMBER_SYSTEM == SIGN_MAGNITUDE
> Null pointer bit pattern == All ones
>
> I'm writing some fully-portable code and would love to actually compile
> it for, and test it on, a very strange machine :-D

Such a machine would be interesting for testing portability. The best
solution is a configurable simulator and compiler that lets you specify
many of these implementation-defined and otherwise variable attributes.

Writing the simulator sounds interesting -- writing a configurable compiler
seems like a lot of work!

Perhaps something based on a C language interpreter would be better. Are
there any C interpreters that would make a good starting point for a
configurable target?

--
Thad

On Tue, 1 Jan 2008 17:11:48 -0600, Tomás Ó hÉilidhe wrote
(in article <Xns9A18EBF80CB1Etoelavabitcom--at--194.125.133.14>):

>
> Has anyone here got access to a ridiculous computer? Something like:
>
> CHAR_BIT == 9
> PADDING_BITS(int) != 0
> NUMBER_SYSTEM == SIGN_MAGNITUDE
> Null pointer bit pattern == All ones
>
> I'm writing some fully-portable code and would love to actually compile it
> for, and test it on, a very strange machine :-D
>
> Or does anyone know of a particular machine I could get my hands on very
> cheaply just for playing around with, something very strange which I can
> get a compliant C89 compiler for? An old supercomputer perhaps? The main
> thing I'd want is a strange value for CHAR_BIT, but I'd be ecstatic if the
> integer types had padding aswell!
>


There was a DS9K for sale on ebay last week, but somebody snapped it
up.

--
Randy Howard (2reply remove FOOBAR)
"The power of accurate observation is called cynicism by those
who have not got it." - George Bernard Shaw

Randy Howard wrote:
> On Tue, 1 Jan 2008 17:11:48 -0600, Tomás Ó hÉilidhe wrote
>>
>> I'm writing some fully-portable code and would love to actually compile it
>> for, and test it on, a very strange machine :-D
>>
>> Or does anyone know of a particular machine I could get my hands on very
>> cheaply just for playing around with, something very strange which I can
>> get a compliant C89 compiler for? An old supercomputer perhaps? The main
>> thing I'd want is a strange value for CHAR_BIT, but I'd be ecstatic if the
>> integer types had padding aswell!
>>
>
> There was a DS9K for sale on ebay last week, but somebody snapped it
> up.

... and left the seller negative feedback, in ones' complement.

More seriously: The IBM AS/400 (iSeries?) is said to handle
pointers in a way that trashes some widely-held but non-portable
assumptions. It doesn't meet all your requirements for strange-
sized characters and so on, but it might be worth while seeing
whether IBM operates a public grid or something on which they'll
rent you some time. (If your stated goal is an AS/400 port of
the software, they might even waive a chunk of the rental.)

--
Eric Sosman
esosman--at--ieee-dot-org.invalid

I came across Univac machines today. Supposedly one of them had 72-Bit
bytes.

I also read that one of their machines consumed 125 kW :-O I plugged
two 2 kilowatt heaters into an extension lead today, and the *extension
lead* got nice and hot, never mind the heaters.

--
Tomás Ó hÉilidhe

"Tomï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï ¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ " wrote:
> I came across Univac machines today. Supposedly one of them had 72-Bit
> bytes.
>
> I also read that one of their machines consumed 125 kW :-O I plugged
> two 2 kilowatt heaters into an extension lead today, and the *extension
> lead* got nice and hot, never mind the heaters.

<offtopic>
Don't Do That. At least in the UK, 4kW through one plug is over 17A of
current (230V voltage), and you shouldn't ever try to draw more than 13A
through a plug. Many house fires have been started in this way.

In other countries, it's still a bad idea, though the details may differ.
</offtopic>

PS: FYI, your name is still screwed in my newsreader.

On Fri, 4 Jan 2008 05:04:18 -0600, Philip Potter wrote
(in article <fll3rj$kh9$1--at--aioe.org>):

> PS: FYI, your name is still screwed in my newsreader.

Get a better one.

--
Randy Howard (2reply remove FOOBAR)
"The power of accurate observation is called cynicism by those
who have not got it." - George Bernard Shaw

Philip Potter <pgp--at--doc.ic.ac.uk> wrote in comp.lang.c:

> Don't Do That. At least in the UK, 4kW through one plug is over 17A of
> current (230V voltage), and you shouldn't ever try to draw more than
> 13A through a plug. Many house fires have been started in this way.

(I find the ambiguous spelling of "lead" irritating so I'm going to
replace it with "led" or "leed" where applicable)

The extension leed has four sockets, each of which leed to a common 13 A
fuse.

I should be able to plug whatever I want in wherever I want... worst case
scenario being a blown fuse.

(I originally set out writing this post to prove you wrong... but I did
the calculations and you turned out to be right)

Power = Voltage * Current
4000 = 230 * Current

Current = 4000 / 230 = 17.4 A

My original argument would have been that I should be able to plug
anything in wherever I want because the plugs all have 13 A fuses, but
now I'm at a loss to explain why a fuse wasn't blown when I was drawing
17.4 amps through one plug... ?

There's only two reasonable conclusions I can draw. Either:

A) The fuse is letting 17 A pass, and so isn't really a 13 A fuse at all.
B) Each heater is in fact only 1.5 kW (or maybe just one of them is
weaker)

Anyway, a question I'd like to ask: Should an actual plug or leed
ever become hot? Because the plastic of the plug from the extension leed
going into the wall was pretty hot, as was the leed.

I'm an electronic engineer myself but I still wouldn't have batted an
eyelid about plugging things in willy-nilly, (even into a mutli-adapter),
because the worst thing that should happen is a blown fuse.

I bought a smoke alarm along with the heaters though, just in case,
and I have it mounted right above them. I've got smoke alarms on the
ceilings of the rooms in my house, but I've also got ones mounted
specifically above high-risk devices (my washing machine, dryer, and the
heaters). I've only ever had one fire in my house and it was in a washing
machine; thankfully it was noticed in time and the unit was dragged out
into the back garden before any real damage was done.

--
Tomás Ó hÉilidhe