Hi Andre,
When I will add support for recursive signal paths, it would be
interesting to try that with a large design. So I am interested by your
suggestion, thanks.
By the way, it could be really helpfull to add an option like the
current --disp-tree option but which would instead print the design
hierarchy in the wave option file format.
Best regards,
Jonas
Le 21. 07. 16 à 22:27, Andre a écrit :
Hi All,
I have an extremely large design which has 64 Pipelined Processors
that are described in a three dimensial array with a transputer like
comms behind each one.
At present I'm getting approx 1us per minute for the simulation.
If you would like a very large design to test out the waves option
then I may be able to help, specifically having have a 3d generate loop.
Whatever happens this is a brilliant tool and it's good to get away
from the free but crippled version of Modelsim I'm using until recently.
Best Regards
Andre'
------------------------------------------------------------------------
*From:* Ghdl-discuss <ghdl-discuss-boun...@gna.org> on behalf of
ghdl-discuss-requ...@gna.org <ghdl-discuss-requ...@gna.org>
*Sent:* 20 July 2016 10:00
*To:* ghdl-discuss@gna.org
*Subject:* Ghdl-discuss Digest, Vol 128, Issue 15
Send Ghdl-discuss mailing list submissions to
ghdl-discuss@gna.org
To subscribe or unsubscribe via the World Wide Web, visit
https://mail.gna.org/listinfo/ghdl-discuss
Ghdl-discuss -- GHDL discuss list - News Info Page
<https://mail.gna.org/listinfo/ghdl-discuss>
mail.gna.org
To see the collection of prior postings to the list, visit the
Ghdl-discuss Archives. Using Ghdl-discuss: To post a message to all
the list members ...
or, via email, send a message with subject or body 'help' to
ghdl-discuss-requ...@gna.org
You can reach the person managing the list at
ghdl-discuss-ow...@gna.org
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Ghdl-discuss digest..."
Today's Topics:
1. Re: Collaboration in the GHDL project (Jonas Baggett)
2. Re: Collaboration in the GHDL project (Tristan Gingold)
3. Enabling Waveform Capture after some time (Thomas Dejanovic)
4. Re: Enabling Waveform Capture after some time (Tristan Gingold)
----------------------------------------------------------------------
Message: 1
Date: Tue, 19 Jul 2016 12:10:18 +0200
From: Jonas Baggett <jon...@tranquille.ch>
To: ghdl-discuss@gna.org
Subject: Re: [Ghdl-discuss] Collaboration in the GHDL project
Message-ID: <21c0eee6-c99f-fe9c-15ad-87b775340...@tranquille.ch>
Content-Type: text/plain; charset="utf-8"; Format="flowed"
Hello Tristan,
Now I have a first working version of signal selection. Currently it
supports only full name signals, meaning that wildcards aren't supported
yet (that's number 3 in your list of important features). And currently
it works only for the ghw wave format, but before I extend it to all
formats, do you prefer that I make a pull request now to let you review
my code or should I first add the support for all formats ? And I still
need to add a help description and the new option I made
(--wave-option-file).
Some open questions :
- Currently when the same signal is written several times in the wave
option file, this is silently ignored. Does it worth it to add a tag to
my tree in order to be able to catch that and make it an error/warning ?
- When the wave option file is void of any signal, then all the signals
will be displayed just like if no wave option file was provided. Should
it instead display no signals at all ?
Regards,
Jonas
Le 09.07.16 à 07:20, Tristan Gingold a écrit :
> On 09/07/16 06:59, Jonas Baggett wrote:
>> Hello Tristan,
>>
>>> On 07/07/16 18:50, Jonas Baggett wrote:
>>>> Hello Tristan,
>>>>
>>>> There is also the case of signals in a package : I would suggest the
>>>> following syntax : pkg.the_signal. It seems better to me than
>>>> /pkg/the_signal.
>>>
>>> Why ? /pkg/the_signal is more regular.
>>> Note you can have two packages with the same name in two different
>>> libraries, so maybe it should be /lib/pkg/the_signal. It is also
>>> possible for an entity to have the name of a library, so maybe it
>>> should be /@lib/pkg/the_signal. At this point we almost follow
>>> the syntax of external path names except '/' is used instead of '.'.
>> Because it's with this synthax that we access the signal in VHDL. So I
>> thought that when someone reads lib.pkg.the_signal, it will be clear to
>> him at first glance that we are accessing a signal inside a package.
>> Nevertheless, with the /@lib/pkg/the_signal synthax, it's also clear.
>> Then if you prefer this synthax, it's fine to me too.
>
> / or . is ok with me. I just didn't understand why package signal
> names were different.
>
>> Now I have finished to do the file parser and the tree builder. I have
>> one question : I tried to use the Types package located in
src/types.ads
>> from my package located in src/grt but I get a compile error as the
>> compiler can't find it because it is at level - 1. Is it possible to
>> achieve that ?
>
> No, it is not possible. The files in src/grt are the one for the
> runtime (used during simulation). They are completely independent
> from compiler files (src and src/vhdl).
>
> I agree it is not very clear as grt/ is within src/. Maybe grt/
> should be moved ?
>
>>>> PS : Good luck for France tonight .
>>>
>>> Usually France looses against Germany
>> Except this time ;).
>
> Surprises happen.
>
> Tristan.
>
>
> _______________________________________________
> Ghdl-discuss mailing list
> Ghdl-discuss@gna.org
> https://mail.gna.org/listinfo/ghdl-discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </public/ghdl-discuss/attachments/20160719/c7cb5b53/attachment.html>
------------------------------
Message: 2
Date: Tue, 19 Jul 2016 18:41:24 +0200
From: Tristan Gingold <tging...@free.fr>
To: ghdl-discuss@gna.org
Subject: Re: [Ghdl-discuss] Collaboration in the GHDL project
Message-ID: <4c5af1f1-aa10-b5d1-41a0-cf770684e...@free.fr>
Content-Type: text/plain; charset=utf-8; format=flowed
Hello,
On 19/07/16 12:10, Jonas Baggett wrote:
> Now I have a first working version of signal selection. Currently it
> supports only full name signals, meaning that wildcards aren't supported
> yet (that's number 3 in your list of important features).
Great!
> And currently
> it works only for the ghw wave format, but before I extend it to all
> formats, do you prefer that I make a pull request now to let you review
> my code or should I first add the support for all formats ?
Why not starting review now ?
> And I still
> need to add a help description and the new option I made
> (--wave-option-file).
>
> Some open questions :
>
> - Currently when the same signal is written several times in the wave
> option file, this is silently ignored. Does it worth it to add a tag to
> my tree in order to be able to catch that and make it an error/warning ?
Not sure. With wildcards duplicate may be more common.
> - When the wave option file is void of any signal, then all the signals
> will be displayed just like if no wave option file was provided. Should
> it instead display no signals at all ?
I think this is ok. Displaying no signals is not very interesting!
Might worth emitting a warning.
Thank you for all your efforts!
Tristan.
------------------------------
Message: 3
Date: Wed, 20 Jul 2016 12:16:24 +1000
From: Thomas Dejanovic <thomas.dejano...@gmail.com>
To: ghdl-discuss@gna.org
Subject: [Ghdl-discuss] Enabling Waveform Capture after some time
Message-ID:
<CAFk9-U0UU5cHF6fSNNZXF9vi-JsLSm7F20TAH2cvqrB=gbp...@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Hi all,
I have a test that fails after @1988277252845fs. It's a randomisation
test so i don;t know how to make it fail sooner. The wave file is 525
Mbytes and GTKWave gets killed by my OS while starting up.
Is there any way to specify a start time for data capture? if not,
could this be added as a feature request?
Best Regards, Thomas D.
--
Sent from Tom's Fortress of Solitude!
------------------------------
Message: 4
Date: Wed, 20 Jul 2016 06:50:17 +0200
From: Tristan Gingold <tging...@free.fr>
To: ghdl-discuss@gna.org
Subject: Re: [Ghdl-discuss] Enabling Waveform Capture after some time
Message-ID: <2763a69e-4fc0-4746-7e5f-0a9abd232...@free.fr>
Content-Type: text/plain; charset=utf-8; format=flowed
Hello,
On 20/07/16 04:16, Thomas Dejanovic wrote:
> I have a test that fails after @1988277252845fs. It's a randomisation
> test so i don;t know how to make it fail sooner. The wave file is 525
> Mbytes and GTKWave gets killed by my OS while starting up.
fst files (--fst=) are much smaller. Did you try them ?
> Is there any way to specify a start time for data capture?
No.
> if not,
> could this be added as a feature request?
Feel free to create an issue on github for this request!
Regards,
Tristan.
------------------------------
Subject: Digest Footer
_______________________________________________
Ghdl-discuss mailing list
Ghdl-discuss@gna.org
https://mail.gna.org/listinfo/ghdl-discuss
------------------------------
End of Ghdl-discuss Digest, Vol 128, Issue 15
*********************************************
_______________________________________________
Ghdl-discuss mailing list
Ghdl-discuss@gna.org
https://mail.gna.org/listinfo/ghdl-discuss
_______________________________________________
Ghdl-discuss mailing list
Ghdl-discuss@gna.org
https://mail.gna.org/listinfo/ghdl-discuss
