Which computer was the first one supported single step execution and
breakpoint? For microprocessor, 4004 didn’t have trap flag or int 3
instruction. 8086 should be the first CPU which had both trap flag and
int 3 (am I wrong?).But how about non-PC computers,does MARK I/II or
Eniac support any debug mechanism?
I have been thinking and googling this question these days. Do you think
this question is meaningful? Any comments or links will be appreciated!
Best Regards
Raymond Zhang
Hello,
* On Fri, Jun 03, 2005 at 01:36:41PM +0800 Zhang, Raymond wrote:
Which computer was the first one supported single step execution and
breakpoint? For microprocessor, 4004 didn’t have trap flag or int 3
instruction. 8086 should be the first CPU which had both trap flag and
int 3 (am I wrong?).
Well, I think this is rather off-topic here, but anyway.
int 3 is not the only solution for breakpoints. 
For example, 6502
had BRK, 6800 had SWI, 8080 had RST, which could all be used (and were
used) for breakpoints.
For single steps, often external interrupt sources were used. For
example, in the case of the 6502, use a 6522 timer to generate an
interrupt just 1 tick after been started. In this case, the 6502
executes exactly one command and returns immediately.
Thus, the 8086 was clearly not the first CPU supporting both. And,
remember, In-Circuit-Emulators (ICE) were available, too.
Regards,
Spiro.
–
Spiro R. Trikaliotis
http://www.trikaliotis.net/
Oh please, here we go again. Some young squirt barely dry behind the ears
comes along and thinks his generation has solved it all. The 4004, 8080,
8086, or 6x00 were not even a gleam in the creators eye, let alone the
creator being a gleam in their father’s eye, when the first program was
single stepped through a computer.
Remember all those flashing lights you see on the main consoles of computers
in documentaries or movies from the 50’s and 60’s? Below most of those
lights were switches used to turn the corresponding bit on or off. Using
other control switches you could set memory locations, change registers or
halt the machine and completely re-write the machine code. I was doing that
in 1974 on a Microdata 1600, years before Wazniak and Jobs, the Imsai, and
the TRS-80. I’ve had to manually key in the boot instructions using the
front panel switches and lights, to read in the paper tape to load the boot
loader from the 8" floppy to then load the OS off of the 14" single platter
10 MB removable harddisk.
To even THINK that the concepts of single stepping and debug waited for the
proper batch of silly-cone to be mixed is arrogance in the extreme. Dorothy
was using them when I was still running around in droopy diapers in the
early 50’s in West Texas. Hell, who do you think first coined the term “bug”
from which “debug” is derived? Truth to tell, it would surprise me not to
learn Babbage had a single step capability. Indeed, single step may have
been his principle mode of operation.
Y’all have good day now, ya hear? 
The personal opinion of
Gary G. Little
“Spiro Trikaliotis” wrote in message
news:xxxxx@windbg…
> Hello,
>
> * On Fri, Jun 03, 2005 at 01:36:41PM +0800 Zhang, Raymond wrote:
>
>> Which computer was the first one supported single step execution and
>> breakpoint? For microprocessor, 4004 didn’t have trap flag or int 3
>> instruction. 8086 should be the first CPU which had both trap flag and
>> int 3 (am I wrong?).
>
> Well, I think this is rather off-topic here, but anyway.
>
> int 3 is not the only solution for breakpoints. For example, 6502
> had BRK, 6800 had SWI, 8080 had RST, which could all be used (and were
> used) for breakpoints.
>
> For single steps, often external interrupt sources were used. For
> example, in the case of the 6502, use a 6522 timer to generate an
> interrupt just 1 tick after been started. In this case, the 6502
> executes exactly one command and returns immediately.
>
> Thus, the 8086 was clearly not the first CPU supporting both. And,
> remember, In-Circuit-Emulators (ICE) were available, too.
>
> Regards,
> Spiro.
>
> –
> Spiro R. Trikaliotis
> http://www.trikaliotis.net/
>
Gary G. Little wrote:
Oh please, here we go again. Some young squirt barely dry behind the ears
comes along and thinks his generation has solved it all. The 4004, 8080,
8086, or 6x00 were not even a gleam in the creators eye, let alone the
creator being a gleam in their father’s eye, when the first program was
single stepped through a computer.Remember all those flashing lights you see on the main consoles of computers
in documentaries or movies from the 50’s and 60’s? Below most of those
lights were switches used to turn the corresponding bit on or off. Using
other control switches you could set memory locations, change registers or
halt the machine and completely re-write the machine code.
Absolutely right. The Control Data 3000 mainframes, circa 1960, had
some very cool single-step and breakpoint facilities on that flashy
front console. The IBM/360 front panel was also very useful for
low-level debugging.
–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.