FPGA Central

Hi,

Does VHDL have a goto statement ?

I would like to use a goto statement to prevent "expensive" procedure calls
etc... (at least in pascal it would be expensive)

But maybe for VHDL it doesn't matter if code is split up in procedures ?

Still... It could be handy to write stuff in pascal with goto's and then
later convert it to VHDL... but since it doesn't have goto's I guess those
goto's will have to be replaced by procedure calls later..

Bye,
Skybuck.

GOTOs inside the same scope can always be replaced by structural
statements (IF, LOOP), not procedure calls. Pascal supports only such
gotos. The efficiency of IF,LOOP is the same as for gotos, so there's
no point.

Hubble.

Skybuck Flying wrote:
> Hi,
>
> Does VHDL have a goto statement ?
>
> I would like to use a goto statement to prevent "expensive" procedure calls
> etc... (at least in pascal it would be expensive)
>
> But maybe for VHDL it doesn't matter if code is split up in procedures ?
>
> Still... It could be handy to write stuff in pascal with goto's and then
> later convert it to VHDL... but since it doesn't have goto's I guess those
> goto's will have to be replaced by procedure calls later..

Ya know, your designs are doomed to fail because you've got a
procedural software mindset.

You're clearly not a hardware designer (as evidenced by your posts on
sci.electronics.design) and you cannot impose a software mindset on
hardware design.

Instead of trying to learn VHDL, you should consider learning logic
design. After you've mastered that, you can then see how you can use
VHDL to Describe your Hardware.

-a

"Andy Peters" <[email protected]> wrote in message
news:[email protected] ps.com...
> Skybuck Flying wrote:
> > Hi,
> >
> > Does VHDL have a goto statement ?
> >
> > I would like to use a goto statement to prevent "expensive" procedure
calls
> > etc... (at least in pascal it would be expensive)
> >
> > But maybe for VHDL it doesn't matter if code is split up in procedures ?
> >
> > Still... It could be handy to write stuff in pascal with goto's and then
> > later convert it to VHDL... but since it doesn't have goto's I guess
those
> > goto's will have to be replaced by procedure calls later..
>
> Ya know, your designs are doomed to fail because you've got a
> procedural software mindset.
>
> You're clearly not a hardware designer (as evidenced by your posts on
> sci.electronics.design) and you cannot impose a software mindset on
> hardware design.
>
> Instead of trying to learn VHDL, you should consider learning logic
> design. After you've mastered that, you can then see how you can use
> VHDL to Describe your Hardware.

Dude get real...

Many, many, many people have programmed cpu's in high level languages...

It's even possible to code a tiny cpu in assembler.

I choose delphi since I am good at it.

And coding a cpu in delphi/pascal where I am good in as already though
enough.

Actually it's not the programming that's though. The programming would be
****ing simple, that's no challenge anymore for me.

The problem is concepts, design, instruction set, shit like that.

And since I am trying to do something which nobody has ever done before I
have to invent it all myself

And that is going to cost lot's of time, and the end result would be what ?
maybe a crappy cpu or maybe a fun cpu or maybe something really
asthnoshing... whatever I won't get rich from it...

It's probably going to be utterly ****ing useless lol...

If I write it in delphi/pascal at least I can use all that stuff in windows
apps

Currently I just need a challenge and this stuff is challenging

Even if I can't do it in VHDL... it's still interesting and fun to see how
it all works etc.

So beat it... tututututututututututut

Logic design ? man what the hell you talking about ? I have programmed for
more than 10 years maannn...

10 years man.. almost non stop man... gimme a break man... you better mean
something else with logic shit man otherwise you aint dope man. I had enough
logic in my ass to last a lifetime lol.

Bye,
Skybuck :P:P:P

"Skybuck Flying" <[email protected]> wrote in message
news:dd877s$tah$[email protected]..
>
> "Andy Peters" <[email protected]> wrote in message
> news:[email protected] ps.com...
>> Skybuck Flying wrote:
>> > Hi,
>> >
>> > Does VHDL have a goto statement ?
>> >
>> > I would like to use a goto statement to prevent "expensive" procedure
> calls
>> > etc... (at least in pascal it would be expensive)
>> >
>> > But maybe for VHDL it doesn't matter if code is split up in procedures
>> > ?
>> >
>> > Still... It could be handy to write stuff in pascal with goto's and
>> > then
>> > later convert it to VHDL... but since it doesn't have goto's I guess
> those
>> > goto's will have to be replaced by procedure calls later..
>>
>> Ya know, your designs are doomed to fail because you've got a
>> procedural software mindset.
>>
>> You're clearly not a hardware designer (as evidenced by your posts on
>> sci.electronics.design) and you cannot impose a software mindset on
>> hardware design.
>>
>> Instead of trying to learn VHDL, you should consider learning logic
>> design. After you've mastered that, you can then see how you can use
>> VHDL to Describe your Hardware.
>
> Dude get real...
>
> Many, many, many people have programmed cpu's in high level languages...
>
> It's even possible to code a tiny cpu in assembler.
>
> I choose delphi since I am good at it.
>
> And coding a cpu in delphi/pascal where I am good in as already though
> enough.
>
> Actually it's not the programming that's though. The programming would be
> ****ing simple, that's no challenge anymore for me.
>
> The problem is concepts, design, instruction set, shit like that.
>
> And since I am trying to do something which nobody has ever done before I
> have to invent it all myself
>
> And that is going to cost lot's of time, and the end result would be what
> ?
> maybe a crappy cpu or maybe a fun cpu or maybe something really
> asthnoshing... whatever I won't get rich from it...
>
> It's probably going to be utterly ****ing useless lol...
>
> If I write it in delphi/pascal at least I can use all that stuff in
> windows
> apps
>
> Currently I just need a challenge and this stuff is challenging
>
> Even if I can't do it in VHDL... it's still interesting and fun to see how
> it all works etc.
>
> So beat it... tututututututututututut
>
> Logic design ? man what the hell you talking about ? I have programmed for
> more than 10 years maannn...
>
> 10 years man.. almost non stop man... gimme a break man... you better mean
> something else with logic shit man otherwise you aint dope man. I had
> enough
> logic in my ass to last a lifetime lol.
>

I think you're missing the point Skybuck.

You have to understand the HARDWARE IMPLICATIONS of what VHDL is trying to
achieve.
You may be an excellent programmer & understand logic concepts (AND, OR
etc), BUT
that is not the same as understanding h/w.
I think I have to agree with Andy Peters on this.

Niv.

Skybuck Flying wrote:

>>Ya know, your designs are doomed to fail because you've got a
>>procedural software mindset.
>>
>>You're clearly not a hardware designer (as evidenced by your posts on
>>sci.electronics.design) and you cannot impose a software mindset on
>>hardware design.


> Dude get real...
>
> Many, many, many people have programmed cpu's in high level languages...
>
> It's even possible to code a tiny cpu in assembler.

Please! This is a VHDL newsgroup. Therefore you get VHDL-related
answers. Programming software for a CPU is really different from
modelling a CPU in hardware.

Modelling hardware is more like e.g. AutoCAD - although it looks like
programming software.



> Logic design ? man what the hell you talking about ? I have programmed for
> more than 10 years maannn...

He is talking about flipflops, latches, combinational gates, tri-state
gates and so on. If you want to model synthesizable circuits, you must
not forget them.

You can use VHDL as a programming language (see Ashenden's VHDL
Cookbook) to _simulate_ some behavior - but remeber, that modelling
hardware is something different. For simulation there may be other
languages or tools, that may fit better. VHDL aims to hardware
description and (additionally) offers simulation features.


Ralf

"Niv" <[email protected]> wrote in message
news:fpNJe.1293$[email protected]..
>
> "Skybuck Flying" <[email protected]> wrote in message
> news:dd877s$tah$[email protected]..
> >
> > "Andy Peters" <[email protected]> wrote in message
> > news:[email protected] ps.com...
> >> Skybuck Flying wrote:
> >> > Hi,
> >> >
> >> > Does VHDL have a goto statement ?
> >> >
> >> > I would like to use a goto statement to prevent "expensive" procedure
> > calls
> >> > etc... (at least in pascal it would be expensive)
> >> >
> >> > But maybe for VHDL it doesn't matter if code is split up in
procedures
> >> > ?
> >> >
> >> > Still... It could be handy to write stuff in pascal with goto's and
> >> > then
> >> > later convert it to VHDL... but since it doesn't have goto's I guess
> > those
> >> > goto's will have to be replaced by procedure calls later..
> >>
> >> Ya know, your designs are doomed to fail because you've got a
> >> procedural software mindset.
> >>
> >> You're clearly not a hardware designer (as evidenced by your posts on
> >> sci.electronics.design) and you cannot impose a software mindset on
> >> hardware design.
> >>
> >> Instead of trying to learn VHDL, you should consider learning logic
> >> design. After you've mastered that, you can then see how you can use
> >> VHDL to Describe your Hardware.
> >
> > Dude get real...
> >
> > Many, many, many people have programmed cpu's in high level languages...
> >
> > It's even possible to code a tiny cpu in assembler.
> >
> > I choose delphi since I am good at it.
> >
> > And coding a cpu in delphi/pascal where I am good in as already though
> > enough.
> >
> > Actually it's not the programming that's though. The programming would
be
> > ****ing simple, that's no challenge anymore for me.
> >
> > The problem is concepts, design, instruction set, shit like that.
> >
> > And since I am trying to do something which nobody has ever done before
I
> > have to invent it all myself
> >
> > And that is going to cost lot's of time, and the end result would be
what
> > ?
> > maybe a crappy cpu or maybe a fun cpu or maybe something really
> > asthnoshing... whatever I won't get rich from it...
> >
> > It's probably going to be utterly ****ing useless lol...
> >
> > If I write it in delphi/pascal at least I can use all that stuff in
> > windows
> > apps
> >
> > Currently I just need a challenge and this stuff is challenging
> >
> > Even if I can't do it in VHDL... it's still interesting and fun to see
how
> > it all works etc.
> >
> > So beat it... tututututututututututut
> >
> > Logic design ? man what the hell you talking about ? I have programmed
for
> > more than 10 years maannn...
> >
> > 10 years man.. almost non stop man... gimme a break man... you better
mean
> > something else with logic shit man otherwise you aint dope man. I had
> > enough
> > logic in my ass to last a lifetime lol.
> >
>
> I think you're missing the point Skybuck.
>
> You have to understand the HARDWARE IMPLICATIONS of what VHDL is trying to
> achieve.
> You may be an excellent programmer & understand logic concepts (AND, OR
> etc), BUT
> that is not the same as understanding h/w.
> I think I have to agree with Andy Peters on this.

Exactly hardware implications have nothing to do with logic design... It's
more about concepts, algorithms, frequencies, bandwidth, caches, ram, rom,
speeds, etc...

Bye,
Skybuck.

Now I think it would be a *very* good idea to stop feeding this troll.

Nicolas

"Skybuck Flying" <[email protected]> wrote in message news:dd6i0q$kh8$[email protected]..
> Hi,
>
> Does VHDL have a goto statement ?
>

No. (thank heaven).

There are 'exit' and 'next' (like break and continue) from a loop,
and there are 'return' statements from subprograms.

These are 'jumps' out of the regular data flow, but nicely constrained and
not a wildcard like a goto.
Also, they are supported by synthesis tools (as well as simulators).

> I would like to use a goto statement to prevent "expensive" procedure calls
> etc... (at least in pascal it would be expensive)

How do you define "expensive" ?
Procedure calls synthesize as much hardware as if you would write them in-line.

>
> But maybe for VHDL it doesn't matter if code is split up in procedures ?

That's right. At least for synthesis it is.
For simulation, probably in-lining is slightly faster.

>
> Still... It could be handy to write stuff in pascal with goto's and then
> later convert it to VHDL... but since it doesn't have goto's I guess those
> goto's will have to be replaced by procedure calls later..
>
> Bye,
> Skybuck.
>
>
>

I think you are a bit rude here, Skybuck....
Anyway, for the difference of programming programs and designing
hardware:
There actually are some approaches to "program" cpu's or hardware...
These can be found in the virtual prototyping world. I worked in a
project doing exactly this, programming models for hardware components
to do design evaluations. Software is quite a lot faster than any
vhdl-simulator and with such an approach you can really evaluate some
aspects of hardware-software-codesign too. The only thing you need is a
good concept of how to handle interfaces between the components and the
timing of your "hardware".
The bad news... Such models will never make real hardware! So there is
not really a point to use vhdl for it. Some object oriented programming
language is much better suited.
If you want hardware, you have to take your model and rewrite it in a
modeling language and here some of your problems may surface for the
first time also.. since now you have to think not only hardware
concepts but also low-level hardware...

Okay.. enough for it...
I hope this gives some ideas..
Cheers,
A.J.