Menu
▾
▴
iverilog-devel
|
From: Martin W. <mai...@ma...> - 2014-12-22 21:46:57
|
I suggested this a while ago, but have now got round to trying it. Rather than
have a single regress.list, I've split it into a number of separate lists:
regress-vlg.list - tests that should work with any standard
Verilog (1364-2005) simulator
regress-sv.list - tests that should work with any standard
SystemVerilog (1800-2012) simulator
regress-vhdl.list - tests that should work with any standard
VHDL simulator
regress-vams.list - tests that should work with any standard
Verilog-AMS simulator
regress-ivl1.list - tests that use Icarus Verilog extensions
or known deviations from the standard,
excluding those in regress-ivl2.list
regress-ivl2.list - tests that rely on lossless expressions
or on the default vvp unformatted real
number display
regress-synth.list - tests that exercise the Icarus Verilog
synthesis functionality
regress-v0.9.list - test exceptions for v0.9
regress-v0.10.list - test exceptions for v0.10
regress-vlog95.list - test exceptions for the vlog95 target
(I may not have put all the tests in their proper categories, but I think it's
mostly right).
The main aim of this change is to reduce the effort needed to maintain the
lists when adding tests that aren't supported by older versions of Icarus or
by the vlog95 target. The scripts only load the test lists that are
appropriate to the compiler version and target being exercised.
I've taken this opportunity to sort the main lists into alphabetic/numeric
order and to make the tabulation more consistent (and hopefully more readable).
I've pushed these new lists, plus two modified scripts, new_vvp_reg.pl and
new_vlog95_reg.pl, to the ivtest master branch. The original regress.list and
perl scripts are still there, unmodified.
Take a look and let me know what you think.
Martin
P.S. I haven't created a test exception list for v0.8 - do we still want/need
to support this?
|
|
From: Stephen W. <st...@ic...> - 2014-12-23 21:59:20
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 This is AWESOME! We've been needing this for a while, and finally one of us has gotten around to it. Good job! On 12/22/2014 01:46 PM, Martin Whitaker wrote: > I suggested this a while ago, but have now got round to trying it. > Rather than have a single regress.list, I've split it into a number > of separate lists: > > regress-vlg.list - tests that should work with any standard > Verilog (1364-2005) simulator regress-sv.list - tests that > should work with any standard SystemVerilog (1800-2012) simulator > regress-vhdl.list - tests that should work with any standard > VHDL simulator regress-vams.list - tests that should work with > any standard Verilog-AMS simulator regress-ivl1.list - tests > that use Icarus Verilog extensions or known deviations from the > standard, excluding those in regress-ivl2.list regress-ivl2.list > - tests that rely on lossless expressions or on the default vvp > unformatted real number display regress-synth.list - tests that > exercise the Icarus Verilog synthesis functionality > regress-v0.9.list - test exceptions for v0.9 regress-v0.10.list > - test exceptions for v0.10 regress-vlog95.list - test exceptions > for the vlog95 target > > (I may not have put all the tests in their proper categories, but I > think it's mostly right). > > The main aim of this change is to reduce the effort needed to > maintain the lists when adding tests that aren't supported by older > versions of Icarus or by the vlog95 target. The scripts only load > the test lists that are appropriate to the compiler version and > target being exercised. > > I've taken this opportunity to sort the main lists into > alphabetic/numeric order and to make the tabulation more consistent > (and hopefully more readable). > > I've pushed these new lists, plus two modified scripts, > new_vvp_reg.pl and new_vlog95_reg.pl, to the ivtest master branch. > The original regress.list and perl scripts are still there, > unmodified. > > Take a look and let me know what you think. > > Martin > > P.S. I haven't created a test exception list for v0.8 - do we still > want/need to support this? > > ------------------------------------------------------------------------------ > > Dive into the World of Parallel Programming! The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot > Media, is your hub for all things parallel software development, > from weekly thought leadership blogs to news, videos, case studies, > tutorials and more. Take a look and join the conversation now. > http://goparallel.sourceforge.net > _______________________________________________ Iverilog-devel > mailing list Ive...@li... > https://lists.sourceforge.net/lists/listinfo/iverilog-devel > - -- Steve Williams "The woods are lovely, dark and deep. steve at icarus.com But I have promises to keep, http://www.icarus.com and lines to code before I sleep, http://www.picturel.com And lines to code before I sleep." -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlSZ5a0ACgkQrPt1Sc2b3iku3QCglETwkn/e8GCS54aDClFGacrO wKoAn165nq8Bnxr0lUFKMt6UuUloCb6h =Dz+J -----END PGP SIGNATURE----- |
|
From: Cary R. <cy...@ya...> - 2014-12-24 01:25:54
|
I have not looked at the actual implementation/changes, but it all sounds like a good idea.
I guess the next big thing as far as the test suite goes is to make it thread aware. I remember someone had pieces of that implemented, but as I remember we actually need something that is a bit more flexible. I believe what we want is a master controller that keeps things in order and then create N threads to do the actual testing where N = the number of processors. The controller is responsible for issuing the threads and caching the results of threads that finish before threads that are still processing earlier tests. This would allow the results file to be generated in a deterministic manner.
I can elaborate more whenever/if someone else wants to take the on.
Cary
On Tuesday, December 23, 2014 1:59 PM, Stephen Williams <st...@ic...> wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
This is AWESOME! We've been needing this for a while, and finally
one of us has gotten around to it. Good job!
On 12/22/2014 01:46 PM, Martin Whitaker wrote:
> I suggested this a while ago, but have now got round to trying it.
> Rather than have a single regress.list, I've split it into a number
> of separate lists:
>
> regress-vlg.list - tests that should work with any standard
> Verilog (1364-2005) simulator regress-sv.list - tests that
> should work with any standard SystemVerilog (1800-2012) simulator
> regress-vhdl.list - tests that should work with any standard
> VHDL simulator regress-vams.list - tests that should work with
> any standard Verilog-AMS simulator regress-ivl1.list - tests
> that use Icarus Verilog extensions or known deviations from the
> standard, excluding those in regress-ivl2.list regress-ivl2.list
> - tests that rely on lossless expressions or on the default vvp
> unformatted real number display regress-synth.list - tests that
> exercise the Icarus Verilog synthesis functionality
> regress-v0.9.list - test exceptions for v0.9 regress-v0.10.list
> - test exceptions for v0.10 regress-vlog95.list - test exceptions
> for the vlog95 target
>
> (I may not have put all the tests in their proper categories, but I
> think it's mostly right).
>
> The main aim of this change is to reduce the effort needed to
> maintain the lists when adding tests that aren't supported by older
> versions of Icarus or by the vlog95 target. The scripts only load
> the test lists that are appropriate to the compiler version and
> target being exercised.
>
> I've taken this opportunity to sort the main lists into
> alphabetic/numeric order and to make the tabulation more consistent
> (and hopefully more readable).
>
> I've pushed these new lists, plus two modified scripts,
> new_vvp_reg.pl and new_vlog95_reg.pl, to the ivtest master branch.
> The original regress.list and perl scripts are still there,
> unmodified.
>
> Take a look and let me know what you think.
>
> Martin
>
> P.S. I haven't created a test exception list for v0.8 - do we still
> want/need to support this?
>
> ------------------------------------------------------------------------------
>
>
Dive into the World of Parallel Programming! The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot
> Media, is your hub for all things parallel software development,
> from weekly thought leadership blogs to news, videos, case studies,
> tutorials and more. Take a look and join the conversation now.
> http://goparallel.sourceforge.net
> _______________________________________________ Iverilog-devel
> mailing list Ive...@li...
> https://lists.sourceforge.net/lists/listinfo/iverilog-devel
>
- --
Steve Williams "The woods are lovely, dark and deep.
steve at icarus.com But I have promises to keep,
http://www.icarus.com and lines to code before I sleep,
http://www.picturel.com And lines to code before I sleep."
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iEYEARECAAYFAlSZ5a0ACgkQrPt1Sc2b3iku3QCglETwkn/e8GCS54aDClFGacrO
wKoAn165nq8Bnxr0lUFKMt6UuUloCb6h
=Dz+J
-----END PGP SIGNATURE-----
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Iverilog-devel mailing list
Ive...@li...
https://lists.sourceforge.net/lists/listinfo/iverilog-devel
|
|
From: Martin W. <mai...@ma...> - 2015-01-05 20:15:41
|
One enthusiastic yes, and no dissenting voices, so I've switched everything over. Shout if anything seems broken. I've preserved the v0.8 exceptions (in regress-v0.8.list) in case anyone wants to resurrect them. Given that we haven't touched v0.8 in a long time, it doesn't seem worth expending much effort on this. Stephen Williams wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > This is AWESOME! We've been needing this for a while, and finally > one of us has gotten around to it. Good job! > > On 12/22/2014 01:46 PM, Martin Whitaker wrote: >> I suggested this a while ago, but have now got round to trying it. >> Rather than have a single regress.list, I've split it into a number >> of separate lists: >> >> regress-vlg.list - tests that should work with any standard >> Verilog (1364-2005) simulator regress-sv.list - tests that >> should work with any standard SystemVerilog (1800-2012) simulator >> regress-vhdl.list - tests that should work with any standard >> VHDL simulator regress-vams.list - tests that should work with >> any standard Verilog-AMS simulator regress-ivl1.list - tests >> that use Icarus Verilog extensions or known deviations from the >> standard, excluding those in regress-ivl2.list regress-ivl2.list >> - tests that rely on lossless expressions or on the default vvp >> unformatted real number display regress-synth.list - tests that >> exercise the Icarus Verilog synthesis functionality >> regress-v0.9.list - test exceptions for v0.9 regress-v0.10.list >> - test exceptions for v0.10 regress-vlog95.list - test exceptions >> for the vlog95 target >> >> (I may not have put all the tests in their proper categories, but I >> think it's mostly right). >> >> The main aim of this change is to reduce the effort needed to >> maintain the lists when adding tests that aren't supported by older >> versions of Icarus or by the vlog95 target. The scripts only load >> the test lists that are appropriate to the compiler version and >> target being exercised. >> >> I've taken this opportunity to sort the main lists into >> alphabetic/numeric order and to make the tabulation more consistent >> (and hopefully more readable). >> >> I've pushed these new lists, plus two modified scripts, >> new_vvp_reg.pl and new_vlog95_reg.pl, to the ivtest master branch. >> The original regress.list and perl scripts are still there, >> unmodified. >> >> Take a look and let me know what you think. >> >> Martin >> >> P.S. I haven't created a test exception list for v0.8 - do we still >> want/need to support this? >> >> ------------------------------------------------------------------------------ >> >> > Dive into the World of Parallel Programming! The Go Parallel Website, >> sponsored by Intel and developed in partnership with Slashdot >> Media, is your hub for all things parallel software development, >> from weekly thought leadership blogs to news, videos, case studies, >> tutorials and more. Take a look and join the conversation now. >> http://goparallel.sourceforge.net >> _______________________________________________ Iverilog-devel >> mailing list Ive...@li... >> https://lists.sourceforge.net/lists/listinfo/iverilog-devel >> > > > - -- > Steve Williams "The woods are lovely, dark and deep. > steve at icarus.com But I have promises to keep, > http://www.icarus.com and lines to code before I sleep, > http://www.picturel.com And lines to code before I sleep." > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v2 > > iEYEARECAAYFAlSZ5a0ACgkQrPt1Sc2b3iku3QCglETwkn/e8GCS54aDClFGacrO > wKoAn165nq8Bnxr0lUFKMt6UuUloCb6h > =Dz+J > -----END PGP SIGNATURE----- > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming! The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net > _______________________________________________ > Iverilog-devel mailing list > Ive...@li... > https://lists.sourceforge.net/lists/listinfo/iverilog-devel > |
|
From: Cary R. <cy...@ya...> - 2015-01-05 22:42:00
|
Yes keep what we have for V0.8 in some manner in case we want to update the test results. The only reason we have kept it as an option is because it is the only branch that supports synthesis. I used to keep the V0.8 results in sync, but at the moment I do not have V0.8 compiled on any machines so it has been neglected for some time.
Cary
On Monday, January 5, 2015 12:16 PM, Martin Whitaker <mai...@ma...> wrote:
One enthusiastic yes, and no dissenting voices, so I've switched everything
over. Shout if anything seems broken.
I've preserved the v0.8 exceptions (in regress-v0.8.list) in case anyone wants
to resurrect them. Given that we haven't touched v0.8 in a long time, it
doesn't seem worth expending much effort on this.
Stephen Williams wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> This is AWESOME! We've been needing this for a while, and finally
> one of us has gotten around to it. Good job!
>
> On 12/22/2014 01:46 PM, Martin Whitaker wrote:
>> I suggested this a while ago, but have now got round to trying it.
>> Rather than have a single regress.list, I've split it into a number
>> of separate lists:
>>
>> regress-vlg.list - tests that should work with any standard
>> Verilog (1364-2005) simulator regress-sv.list - tests that
>> should work with any standard SystemVerilog (1800-2012) simulator
>> regress-vhdl.list - tests that should work with any standard
>> VHDL simulator regress-vams.list - tests that should work with
>> any standard Verilog-AMS simulator regress-ivl1.list - tests
>> that use Icarus Verilog extensions or known deviations from the
>> standard, excluding those in regress-ivl2.list regress-ivl2.list
>> - tests that rely on lossless expressions or on the default vvp
>> unformatted real number display regress-synth.list - tests that
>> exercise the Icarus Verilog synthesis functionality
>> regress-v0.9.list - test exceptions for v0.9 regress-v0.10.list
>> - test exceptions for v0.10 regress-vlog95.list - test exceptions
>> for the vlog95 target
>>
>> (I may not have put all the tests in their proper categories, but I
>> think it's mostly right).
>>
>> The main aim of this change is to reduce the effort needed to
>> maintain the lists when adding tests that aren't supported by older
>> versions of Icarus or by the vlog95 target. The scripts only load
>> the test lists that are appropriate to the compiler version and
>> target being exercised.
>>
>> I've taken this opportunity to sort the main lists into
>> alphabetic/numeric order and to make the tabulation more consistent
>> (and hopefully more readable).
>>
>> I've pushed these new lists, plus two modified scripts,
>> new_vvp_reg.pl and new_vlog95_reg.pl, to the ivtest master branch.
>> The original regress.list and perl scripts are still there,
>> unmodified.
>>
>> Take a look and let me know what you think.
>>
>> Martin
>>
>> P.S. I haven't created a test exception list for v0.8 - do we still
>> want/need to support this?
>>
>> ------------------------------------------------------------------------------
>>
>>
> Dive into the World of Parallel Programming! The Go Parallel Website,
>> sponsored by Intel and developed in partnership with Slashdot
>> Media, is your hub for all things parallel software development,
>> from weekly thought leadership blogs to news, videos, case studies,
>> tutorials and more. Take a look and join the conversation now.
>> http://goparallel.sourceforge.net
>> _______________________________________________ Iverilog-devel
>> mailing list Ive...@li...
>> https://lists.sourceforge.net/lists/listinfo/iverilog-devel
>>
>
>
> - --
> Steve Williams "The woods are lovely, dark and deep.
> steve at icarus.com But I have promises to keep,
> http://www.icarus.com and lines to code before I sleep,
> http://www.picturel.com And lines to code before I sleep."
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2
>
> iEYEARECAAYFAlSZ5a0ACgkQrPt1Sc2b3iku3QCglETwkn/e8GCS54aDClFGacrO
> wKoAn165nq8Bnxr0lUFKMt6UuUloCb6h
> =Dz+J
> -----END PGP SIGNATURE-----
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming! The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net
> _______________________________________________
> Iverilog-devel mailing list
> Ive...@li...
> https://lists.sourceforge.net/lists/listinfo/iverilog-devel
>
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Iverilog-devel mailing list
Ive...@li...
https://lists.sourceforge.net/lists/listinfo/iverilog-devel
|
Thanks for helping keep SourceForge clean.
X