CoWare

Ajit, you make a good point about writing models but in some cases, that may only apply to new components or new systems. The Synopsys virtual platform 'trio' can also tackle the issue but taking advantage of reusable IP in TLM and/or RTL ... or even make use of Carbon to generate C/C++ models from existing RTL IP. Synopsys and CoWare have been active contributing members of the SPIRIT Consortium so they are aware of making use of IP-XACT to quickly intake reusable IP if that support exists. (Virtutech is listed as a reviewing member but I'm surprised that VaST was not; maybe my memory is bad but I was once active in SPIRIT and thought they were present in that community ... a moot point, now!) I don't see ARM Fast Models nor Carbon as being out of the picture with that Synopsys trio.

Synopsys buying 3 virtual platforms doesn't solve their fundamental deployment problem. Virtio, VAST and CoWare are all based on a services model, where they need to hand write models for every part of the platform they don't have a model. Sorry to say that in practice, this can take months. To date the only scalable virtual platform I have seen that you can deploy within a couple weeks is a combination of using ARM's Fast Models along with Carbon's Model Studio (for model generation of other existing IP).

I see a lot of blogs that are talking about how this affects the technology development and market, in general. Remember, all these products- VAST, Coware and Virtio are completely different. Is Synopsys going to support three platforms for the same end customer? Any ideas as to which one will be dropped. Mirabilis Design is probably the only free-standing System-level companies remaining. We are in the system sizing and project feasibility arena- not really in software virtual prototyping. We use Qemu as our virtual prototyping add-on. it is indeed a great product. The announcement of Synopsis purchasing CoWare/VAST is truly good news for the whole industry. This acquisition validates System Level Design today and more importantly, Architectural Design, moving forward. With chips and systems becoming more complex, the need to understand system sizing, transaction analysis, power, performance, etc. coupled with fast construction of models has driven Mirabilis sales numbers to exceed 30% growth in 2009. This acquisition makes clear what we have been seeing for the past two years from our customers. The whole architectural area is exploding as design teams are finally realizing the importance of understanding the application from an architectural standpoint far in advance of chip design or even platform layout. It is interesting to note, five years ago both Synopsys and Cadence moved away from System Level Design by selling off or ending products in this space. This recent acquisition is a positive sign that the major players in EDA are moving back into the System space. We are not going to be the only ones talking Architectural Design. Mirabilis Design is uniquely positioned in the System / Architectural space and this confirms our strategy.

Interesting times.. Just a point: It is not strictly correct to brand qemu as "ISS".. Qemu can also do system simulation for a variety of platforms.

Paul, it would be interesting if Cadence were to just take the money and run on CoWare's 'exit' by this Synopsys acquisition. Will they take the cash or take back SPW (and BoNES)? If they only take the cash, then SPW will have belonged to Cadence (after purchase from Comdisco) and then Synopsys. I wonder how many other EDA platforms changed ownership between Big 3 EDA firms? Given that Brent Favre is a Minnesta Viking, anything's possible, I guess! Also, Synopsys is now faced with resolving a portfolio of VaST, CoWare, Virtio and their own Innovator virtual platform tool. But what will Cadence do? Is there something to acquire, or purchase like ARM's SoC Designer (Axys) now under Carbon's umbrella?