UVM and OVM future

Hi all,
I have heard about UVM and I have some questions regarding it. In particular:

  • UVM will be fully backward compatible with OVM, from functionality and naming convention point of view?
  • Any hints about UVM availability? Are there any plans to submit UVM to IEEE?
  • Anyone have an idea on how the OVM community will be able in the future to contribute to UVM?
  • What is going to happen to ovmworld site once UVM methodology will be delivered? There is the intention to maintain this site or close it?

Thanks and best regards
Alberto

Hello Alberto,

These are some great questions. While the Accellera VIP TSC (technical subcommittee) is just getting started with UVM, let me take a stab at sharing my view on these questions.

  • UVM full backward compatibility

    • The committee did select OVM as the base for UVM which is the first step towards backward compatibility with OVM. I believe that it is important for users to voice their strong opinions in the committee. Certainly maintaining OVM type names and use models would enhance backward compatibility so it would be great for you or your company to make that requirement known to the TSC.
  • UVM availability and IEEE standardization

    • The committee vote to have a version of UVM available by the end of Q1 2010 so it will need to decide what content will be available in that release in short order. Regarding IEEE standardization the committee will have to vote on that.
  • OVM community contributions to the UVM

    • With OVM as the base of UVM, we expect the OVM community will have a lot of influence on the direction of the UVM. With more than 20K downloads of the 38 contributions so far, it is a proven path to add new technology to the OVM. The VIP TSC did vote to add new technology to the OVM base as well, and the contributions were named as an important source for that new technology.
  • OVM World in a UVM World

    • We do expect OVM World to continue for quite some time. UVM will be functional on day 1, whenever that day occurs, because it is based on OVM. However, there are 10K+ members of the OVM world and thousands of active projects. The community will need to support itself even as we welcome new UVM users. For sure, the next few months will be very interesting!

Thanks for the questions and keep them coming!

Umer

Hi Umer

Thanks for the prompt reply

If the effort of both Cadence & Mentor is going in the UVM direction for me it is definitively a must to have something that is fully backward compatible to OVM
What I am happy to skip is restarting again to restructure my VIPs; for me are enough the changes I experiment across OVM releases ;-)

Thanks again
Regards
Alberto

hi all,

Does this UVM, a combined effort of both the VMM + OVM methodology as going to be proposed by Accellera or its completely a different methodology?? Can any one brief it out plz!!

Thanks in Adv,

So far, all that has been resolved is a series of ballots; no code has been reviewed. Here are some of those resolutions relevant to the question:

  • Resolved: The OVM version 2.0.3 will be used as the starting point for the UVM. 12 yes, 2 no, motion passes
  • Resolved: The name of the CBCL, the specification document, and API will be the “Universal Verification Methodology” and will use the acronym “UVM”. 14 yes, 0 no, motion passes
  • Resolved: Complementary functionality from the VMM, OVM, and/or other contributions from Accellera VIP TSC members may be incorporated into the UVM subject to the vote of the Accellera VIP TSC, 12 yes, 2 no, motion passes
  • Resolved: The goal of the Accellera VIP TSC is to have the first cut release of the UVM done by Q1’10. 12 yes, 1 no, motion passes

A “starting point” in no way guarantees “backward compatibility”. However, 11 out of the 14 voting members are users. I would assume those users are going to put the pressure on to make things as compatible as possible.

Dave Rich