Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
search some good vhdl preprocessor (opensource)
204 views
Skip to first unread message
troll greenЪ
Nov 8, 2012, 11:06:45 AM11/8/12
to
Can anyone help?
Main tags: opensource, good architectured.
Google, opencores didn't help :)
Thanks!
Main tags: opensource, good architectured.
Google, opencores didn't help :)
Thanks!
HT-Lab
Nov 8, 2012, 11:26:34 AM11/8/12
to
remember what I changed, anyway you can download the source
http://sourceforge.net/projects/vhdlpp/
Hans
www.ht-lab.com
Rob Gaddi
Nov 8, 2012, 11:49:06 AM11/8/12
to
preprocessors in the past. Inevitably, they allowed me to write
non-standard code that was more difficult to maintain and work with in
the long run.
VHDL is a fantastically extensible language. Whatever it is you think
you need a preprocessor for, you're probably wrong.
--
Rob Gaddi, Highland Technology -- www.highlandtechnology.com
Email address domain is currently out of order. See above to fix.
troll greenЪ
Nov 8, 2012, 12:01:52 PM11/8/12
to han...@htminuslab.com
четверг, 8 ноября 2012 г., 20:26:36 UTC+4 пользователь HT-Lab написал:
Thanks! I've tried it today but I'll try to find something more complex. Looks like https://metacpan.org/release/Hardware-Vhdl-Automake is a good stuff.
https://metacpan.org/source/MYKL/Hardware-Vhdl-Automake-1.00/lib/Hardware/Vhdl/Automake/PreProcessor.pm
https://metacpan.org/source/MYKL/Hardware-Vhdl-Automake-1.00/lib/Hardware/Vhdl/Automake/PreProcessor.pm
troll greenЪ
Nov 8, 2012, 12:10:06 PM11/8/12
to
> I've used cpp,
You mean KPP -- http://klabs.org/richcontent/software_content/kpp.htm
Or i'm wrong?
>M4,
Hard to find - can you provide some link?
>Inevitably, they allowed me to write
> non-standard code that was more difficult to maintain and work with in
> the long run.
Yes...
> Whatever it is you think you need a preprocessor for, you're probably wrong.
I need it to control matlab->HDL flow easily.
F.E. change vhdl coder output parts for using some FPGA embedded DSP blocks.
etc.
Have my own samples, but do not want to reinvent the wheel in the case of bone preprocessor :)
You mean KPP -- http://klabs.org/richcontent/software_content/kpp.htm
Or i'm wrong?
>M4,
Hard to find - can you provide some link?
>Inevitably, they allowed me to write
> non-standard code that was more difficult to maintain and work with in
> the long run.
> Whatever it is you think you need a preprocessor for, you're probably wrong.
F.E. change vhdl coder output parts for using some FPGA embedded DSP blocks.
etc.
Have my own samples, but do not want to reinvent the wheel in the case of bone preprocessor :)
HT-Lab
Nov 8, 2012, 12:22:39 PM11/8/12
to
On 08/11/2012 16:49, Rob Gaddi wrote:
> On Thu, 8 Nov 2012 08:06:45 -0800 (PST)
> troll greenЪ <xorn...@gmail.com> wrote:
>
>> Can anyone help?
>> Main tags: opensource, good architectured.
>>
>> Google, opencores didn't help :)
>> Thanks!
>
> I've used cpp, M4, and my own proprietary code all as VHDL
> preprocessors in the past. Inevitably, they allowed me to write
> non-standard code that was more difficult to maintain and work with in
> the long run.
>
> VHDL is a fantastically extensible language. Whatever it is you think
> you need a preprocessor for, you're probably wrong.
Probably he isn't as preprocessing is being discussed in the P1076
> On Thu, 8 Nov 2012 08:06:45 -0800 (PST)
> troll greenЪ <xorn...@gmail.com> wrote:
>
>> Can anyone help?
>> Main tags: opensource, good architectured.
>>
>> Google, opencores didn't help :)
>> Thanks!
>
> I've used cpp, M4, and my own proprietary code all as VHDL
> preprocessors in the past. Inevitably, they allowed me to write
> non-standard code that was more difficult to maintain and work with in
> the long run.
>
> VHDL is a fantastically extensible language. Whatever it is you think
> you need a preprocessor for, you're probably wrong.
steering group:
http://www.eda.org/twiki/bin/view.cgi/P1076/VHDLPreprocessor
I do hope we don't end up with the unreadable ' or is it the ` character...
Hans
www.ht-lab.com
>
troll greenЪ
Nov 8, 2012, 12:42:46 PM11/8/12
to
>I do hope we don't end up with the unreadable ' or is it the ` character...
Rob Gaddi
Nov 8, 2012, 12:46:24 PM11/8/12
to
On Thu, 8 Nov 2012 09:10:06 -0800 (PST)
troll greenЪ <xorn...@gmail.com> wrote:
> > I've used cpp,
> You mean KPP -- http://klabs.org/richcontent/software_content/kpp.htm
> Or i'm wrong?
>
No, cpp, the C preprocessor. #define, #include, etc. If you've got a
troll greenЪ <xorn...@gmail.com> wrote:
> > I've used cpp,
> You mean KPP -- http://klabs.org/richcontent/software_content/kpp.htm
> Or i'm wrong?
>
standard GNU GCC installation, the preprocessor is a standalone
program, and can be used on anything you like not just C. Good:
super-easy to work with and you already know the syntax. Bad: pretty
limited in terms of what it can do.
>
> >M4,
> Hard to find - can you provide some link?
>
Good: Super powerful. Turing complete in fact; you can make M4 do
pretty much anything you want it to. Bad: Expect to spend time syntax
tuning. Lots and lots of time. One thing that helps with that is to
use whichever command line setting changes the command names from
things like divert and index to m4_divert and m4_index, you'll get far
fewer collisions that way.
> >Inevitably, they allowed me to write
> > non-standard code that was more difficult to maintain and work with in
> > the long run.
> Yes...
>
> > Whatever it is you think you need a preprocessor for, you're probably wrong.
> I need it to control matlab->HDL flow easily.
> F.E. change vhdl coder output parts for using some FPGA embedded DSP blocks.
> etc.
> Have my own samples, but do not want to reinvent the wheel in the case of bone preprocessor :)
std.textio to read those text files from VHDL to configure things.
That's generally kosher for synthesis so long as your use of textio
only happens at elaboration time. Generally you'd wrap a function
around the textio calls, and have that function return something that
can initialize a constant, that can then make the rest of your code do
what it ought. I've used this technique primarily for initializing
ROMs, filter coefficients, etc.
Or have Matlab write out functions, procedures, or
constants to a VHDL package file, and then use that package in your
main code. I've used this technique to create packages that hold
things like Subversion revision, timestamp, etc, in such a way that
they can be synthesized in.
troll greenЪ
Nov 8, 2012, 1:47:00 PM11/8/12
to
> No, cpp, the C preprocessor.
oh, ok :) I forgot about original variant
> Good:
> super-easy to work with and you already know the syntax. Bad: pretty
> limited in terms of what it can do.
> http://www.gnu.org/software/m4/
Thanks, never heard about it before
> Bad: Expect to spend time syntax tuning. Lots and lots of time.
>std.textio to read those text files from VHDL to configure things
Last seen GNU ADA gnatprep preprocessor is very simple and ugly, so
I'll try your naturally VHDL variant :)
Big tnx!
> I've used this technique to create packages that hold
> things like Subversion revision, timestamp, etc
And something like Apache Ant -- http://ant.apache.org/ -- upper.
With simple AntNest GUI -- http://antsnest.sourceforge.net/
:)
Andy
Nov 8, 2012, 2:26:50 PM11/8/12
to han...@htminuslab.com
Hopefully, the only "preprocessing" allowed will be to address tool incompatibility, especially WRT new LRM features not yet universally supported.
I adamantly oppose standardizing anything approaching the common pre-processors in use with other languages today. They are too often abused to allow lazy, unmaintaibable coding, while slowing the development of language features to accomplish the same functionality. Exhibit A: SystemVerilog and UVM.
For those that want to incorporate a pre-processor in their tool flow, that is fine, but as far as the standard is concerned, such capability should be considerably limited.
Just my $0.02...
Andy
I adamantly oppose standardizing anything approaching the common pre-processors in use with other languages today. They are too often abused to allow lazy, unmaintaibable coding, while slowing the development of language features to accomplish the same functionality. Exhibit A: SystemVerilog and UVM.
For those that want to incorporate a pre-processor in their tool flow, that is fine, but as far as the standard is concerned, such capability should be considerably limited.
Just my $0.02...
Andy
Walter
Nov 12, 2012, 4:30:48 PM11/12/12
to
My apologies, I write VHDL code since long time; I never need a preprocessor. You coming from software or has formed in some language with cuisine "C" ?
troll greenЪ
Nov 16, 2012, 12:32:49 PM11/16/12
to
понедельник, 12 ноября 2012 г., 23:30:49 UTC+2 пользователь Walter написал:
Preprocessor - for some special stuff - Matlab->HDL-coder->VDHL-preprocessor (insert some hardware modules optimizations, etc.).
At now playing with
> My apologies, I write VHDL code since long time; I never need a preprocessor. You coming from software or has formed in some language with cuisine "C" ?
>
Good day,
>
Preprocessor - for some special stuff - Matlab->HDL-coder->VDHL-preprocessor (insert some hardware modules optimizations, etc.).
At now playing with
"std.textio to read those text files from VHDL to configure things"
by Rob Gaddi answer.
troll greenЪ
Jan 21, 2013, 7:11:07 PM1/21/13
to
четверг, 8 ноября 2012 г., 19:46:24 UTC+2 пользователь Rob Gaddi написал:
All above is very good and practical way.
Decided to use this after a long period of trying.
When I get more experience and practical examples - try to open some sources and examples.
Thanks for all!
>
> Another option would be to have Matlab write out text files, and use
>
> std.textio to read those text files from VHDL to configure things.
>
> That's generally kosher for synthesis so long as your use of textio
>
> only happens at elaboration time. Generally you'd wrap a function
>
> around the textio calls, and have that function return something that
>
> can initialize a constant, that can then make the rest of your code do
>
> what it ought. I've used this technique primarily for initializing
>
> ROMs, filter coefficients, etc.
>
>
>
> Or have Matlab write out functions, procedures, or
>
> constants to a VHDL package file, and then use that package in your
>
> main code. I've used this technique to create packages that hold
>
> things like Subversion revision, timestamp, etc, in such a way that
>
> they can be synthesized in.
>
----------------------------------------------------------------
> Another option would be to have Matlab write out text files, and use
>
> std.textio to read those text files from VHDL to configure things.
>
> That's generally kosher for synthesis so long as your use of textio
>
> only happens at elaboration time. Generally you'd wrap a function
>
> around the textio calls, and have that function return something that
>
> can initialize a constant, that can then make the rest of your code do
>
> what it ought. I've used this technique primarily for initializing
>
> ROMs, filter coefficients, etc.
>
>
>
> Or have Matlab write out functions, procedures, or
>
> constants to a VHDL package file, and then use that package in your
>
> main code. I've used this technique to create packages that hold
>
> things like Subversion revision, timestamp, etc, in such a way that
>
> they can be synthesized in.
>
All above is very good and practical way.
Decided to use this after a long period of trying.
When I get more experience and practical examples - try to open some sources and examples.
Thanks for all!
0 new messages