Synopsys's VMM and Mentor's AVM

[Davy]

Hi all,

I want to use SystemVerilog to construct next generation of my
testbench.

And I found Synopsys provide VMM while Mentor provide AVM. Anyone can
give some comment on these two methodology? Or are they similar?

I don't know if Synopsys's VMM is open document and open source code.

The AVM cookbook/source code, you can download a free copy from:
http://www.mentor.com/products/fv/_3b715c/cb_dll.cfm

Best regards,
Davy

 

[EdA]

Hi all,

I want to use SystemVerilog to construct next generation of my
testbench.

And I found Synopsys provide VMM while Mentor provide AVM. Anyone can
give some comment on these two methodology? Or are they similar?

I don't know if Synopsys's VMM is open document and open source code.

The AVM cookbook/source code, you can download a free copy from:
http://www.mentor.com/products/fv/_3b715c/cb_dll.cfm

Best regards,
Davy

Davy, sorry for the non-answer, but you may get better results with
that question at
http://www.verificationguild.com/

/Ed

 

[Paul Uiterlinden]

Hi all,

I want to use SystemVerilog to construct next generation of my
testbench.

And I found Synopsys provide VMM while Mentor provide AVM. Anyone
can give some comment on these two methodology? Or are they similar?

A comparison of AVM and VMM in Verification Horizons:
http://lyris.mentor-info.com/t/5001/4363723/7325/1846/
PDF version:
http://lyris.mentor-info.com/t/5001/4363723/7330/1851/

Four articles in the EETIMES on the SystemVerilog reference
verification methodology:
http://www.eetimes.com/news/design/showArticle.jhtml?articleID=183702807
http://www.eetimes.com/news/design/showArticle.jhtml?articleID=187001913
http://www.eetimes.com/news/design/showArticle.jhtml?articleID=188703275
http://www.eetimes.com/news/design/showArticle.jhtml?articleID=192501500

The first is by Mentor, the second by Synopsys (IIRC). Both of course
with their own biases.

I don't know if Synopsys's VMM is open document and open source
code.

As far as I know the VMM book is not an open document:
http://www.vmm-sv.com/

The AVM cookbook clearly is.

The same goes for VMM and AVM itself. AVM is opensource, VMM source is
heavily licensed (word choice from Verification Horizons).

 

[Davy]

Hi Paul,

Thanks a lot!
I also want to know does Cadence provide such verification methodology
like Synopsys and Mentor.

And what's Synopsys (IIRC)'s IIRC mean?

Best regards,
Davy

 

[Ajeetha]

Hi,

As far as I know the VMM book is not an open document:
http://www.vmm-sv.com/

I don't understand this - perhaps you are mixing "open" with "free"?
VMM is a published book so why is it not open? Infact we wrote a book
on "pragmatic approach to VMM adoption" based on that book. (See
www.systemverilog.us if interested). BTW, VMM also ships under
$VCS_HOME/doc.

The AVM cookbook clearly is.

The same goes for VMM and AVM itself. AVM is opensource, VMM source is
heavily licensed (word choice from Verification Horizons).

Quoting from:

http://www.synopsys.com/news/announce/press2005/snps_sourcode_licsvpr.html

SNPS gives source code to VCS users if they request for the same.

Now having said all this, given the status of SV implementation by
major eda vendors, neither VMM nor AVM is truly "portable" as of today
- tools support different subsets just to fit into their individual
methodology, perhaps the tool development was driven by the methodology
team. So when 100% SV implementation is available across vendors, users
may not have an issue of AVM vs. VMM as both will work in any
simulator.

Now, I'm teaching myself AVM and am finding it quite similar to VMM.
Sure VMM has much more stuff, also maturity (given their RVM legacy),
AVM has some "new" concepts such as analysis ports etc. I asked Mentor
if I can openly debate on AVM, no reply yet...

Regards
Ajeetha, CVC
www.noveldv.com

 

[Paul Uiterlinden]

Hi Paul,

Thanks a lot!
I also want to know does Cadence provide such verification
methodology like Synopsys and Mentor.

Not that I know of. I haven't come across it, but I haven't actively
been searching for it.

And what's Synopsys (IIRC)'s IIRC mean?

IIRC is just an abbreviation for "If I remember correctly". Sorry for
the confusion.

 

[Paul Uiterlinden]

Hi,




I don't understand this - perhaps you are mixing "open" with
"free"?

Yes, I am. What I meant it is not freely down-loadable.

VMM is a published book so why is it not open? Infact we wrote a
book on "pragmatic approach to VMM adoption" based on that book.
(See www.systemverilog.us if interested). BTW, VMM also ships under
$VCS_HOME/doc.


Quoting from:

http://www.synopsys.com/news/announce/press2005/snps_sourcode_licsvpr.html

SNPS gives source code to VCS users if they request for the same.

"VCS customers may license the source code at no additional cost to
gain insight into the implementation details."

So that's for VCS users. I'm not a VCS user, so I do not have access
to the source code. In that way it _is_ licensed. The AVM source is
not.

Now having said all this, given the status of SV implementation by
major eda vendors, neither VMM nor AVM is truly "portable" as of
today - tools support different subsets just to fit into their
individual methodology, perhaps the tool development was driven by
the methodology team. So when 100% SV implementation is available
across vendors, users may not have an issue of AVM vs. VMM as both
will work in any simulator.

Now, I'm teaching myself AVM and am finding it quite similar to VMM.
Sure VMM has much more stuff, also maturity (given their RVM
legacy), AVM has some "new" concepts such as analysis ports etc.

I cannot comment on this. I haven't used or really studied neither of
them. I did find this quote from Verification Horizons quite potent:

"(...) I mention this story because it is similar to the thought
process that many of you may be going through in trying to decide how
to adopt a new verification methodology. You have a similar choice to
make - should you go with the AVM or take a look at the VMM? In my
minivan story, the VMM is the used car since it?s really based on old
technology, having simply been ported from OpenVera® to
SystemVerilog. The AVM is the new topof-the-line car that gives you
all of the latest features and the power and flexibility that you
need. Plus your tool and legacy investments are protected because it
is based on an open standard. Which would you rather use to carry
your precious cargo?"

Granted, this is by Mentor Graphics, so perhaps should be taken with a
grain of salt. I really do not have the knowledge to put a value on
quotes likes this. Perhaps you would like to comment on this.

I asked Mentor if I can openly debate on AVM, no reply yet...

Call me naive, but I do not see why such a debate should not be
possible. It's a free world, "free" as in "free speech", not as in
"free bear".

 

[Davy]

Hi,




I don't understand this - perhaps you are mixing "open" with "free"?
VMM is a published book so why is it not open? Infact we wrote a book
on "pragmatic approach to VMM adoption" based on that book. (See
www.systemverilog.us if interested). BTW, VMM also ships under
$VCS_HOME/doc.


Quoting from:

http://www.synopsys.com/news/announce/press2005/snps_sourcode_licsvpr.html

SNPS gives source code to VCS users if they request for the same.

Now having said all this, given the status of SV implementation by
major eda vendors, neither VMM nor AVM is truly "portable" as of today
- tools support different subsets just to fit into their individual
methodology, perhaps the tool development was driven by the methodology
team. So when 100% SV implementation is available across vendors, users
may not have an issue of AVM vs. VMM as both will work in any
simulator.

Now, I'm teaching myself AVM and am finding it quite similar to VMM.
Sure VMM has much more stuff, also maturity (given their RVM legacy),
AVM has some "new" concepts such as analysis ports etc. I asked Mentor
if I can openly debate on AVM, no reply yet...

[snip]

Hi Ajeetha,

Thanks for the explanation.
I think a lot of people will be interested in your openly comment on
both VMM and AVM without biased opinion.
For AVM, it use Apache licence. Is this licence forbid openly debate

Best regards,
Davy

Regards
Ajeetha, CVC
www.noveldv.com

 

[Ajeetha]

Hi Davy,

> Hi Ajeetha,

Thanks for the explanation.
I think a lot of people will be interested in your openly comment on
both VMM and AVM

That's one of the reasons for me to hold back As I have to be
100% correct else Mentor and/or SNPS folks will start pin pointing
errors with my analysis. I would take some more time, but from quick
analysis so far, both are very similar in concepts and I am even
considering an "adaptor" for AVM users to VMM and vice versa - but all
in thoughts, depends on market.

For AVM, it use Apache licence. Is this licence forbid openly debate

Honestly speaking I have NOT read through the license in full and am
not a lawyer either. Hence I would be glad if someone clearly says "yes
we can debate on it". Being an independent consutlant I want to be
friendly to all vendors.

Regards
Ajeetha, CVC
www.noveldv.com

 

[Davy]

Hi Davy,

That's one of the reasons for me to hold back As I have to be
100% correct else Mentor and/or SNPS folks will start pin pointing
errors with my analysis. I would take some more time, but from quick
analysis so far, both are very similar in concepts and I am even
considering an "adaptor" for AVM users to VMM and vice versa - but all
in thoughts, depends on market.




Honestly speaking I have NOT read through the license in full and am
not a lawyer either. Hence I would be glad if someone clearly says "yes
we can debate on it". Being an independent consutlant I want to be
friendly to all vendors.

[snip]
Hi Ajeetha,

I agree with you. And I have sent a mail to apache.org (Apache License
V2 owner) to ask the problem. I will give out the result when got a
replay.

Best regards,
Davy

 

[Adam_Rose]

Honestly speaking I have NOT read through the license in full and am
not a lawyer either. Hence I would be glad if someone clearly says "yes
we can debate on it". Being an independent consutlant I want to be
friendly to all vendors.

[snip]

Mentor Graphics encourages open discussion of anything we have
published at www.mentor.com/go/cookbook. This is why we published it in
that form.

Any code or pseudo code that uses the AVM is fine. Questions refering
to specific items in the documentation are fine.

If you're going to post the contents of the library itself you already
have permission under the license to do so provided you also copy the
header at the top of the file which includes the copyright notice -
although I would have thought this would be unecessary most of the
time.

Adam.

Adam Rose
Verification Technologist
Mentor Graphics.

 

[googler]

I also want to know does Cadence provide such verification methodology
like Synopsys and Mentor.

From what I found from this thread at

http://verificationguild.com/module...opic&t=1404&postdays=0&postorder=asc&start=15
(check the last two posts), Cadence seems to have a verification
methodology called UVM..