Synopsys’ possible acquisition of emulation specialist EVE has been watercooler talk for quite some time. Now it’s happened. It makes the company the first of the big EDA vendors to offer products across all the prototyping and verification steps that are wholly in-house developed. It also reflects the shift away from ‘either/or’ thinking across that part of the flow.
Largely for reasons of cost, time and expertise, some design managers have been wary of not just emulation (and yes, those boxes do cost a fair few spondooliks) but also virtual prototyping (“Are the models good enough?” “System-level – oooh sounds tricky”). They have nevertheless bemoaned the time taken for simulation (“I’m not slacking,” the mallrat t-shirt says. “My simulation’s running.”), and cursed the effort needed to partition an FPGA prototype accurately (“So, why d’you get the Tim Burton haircut, dude?”). Vendors, for their part, have tended to promote the weapons in their arsenals at the expense of others. No surprise there – could you blame them?
The result, until recently, was that rather dogmatic “that’s how it’s done here” approaches to prototyping have been common with perhaps only a couple of techniques used, even as tool and IP vendors worked hard to overcome objections. Take virtual prototyping. Today there are robust, reusable and battle-hardened models as well as much easier-to-use graphical tools available from all the major players. FPGA prototyping and particularly the thorny partitioning issue is constantly innovating.
And then in emulation, as well as acceleration advances there’s been a solid cost-down drive with steps towards concepts such as virtualized emulation that allow a box to be used across geographically spread sites on the same design team (indeed, we’ll be running our interview with Mentor Graphics’ CEO Wally Rhines later this week where he discusses the path toward emulation for the masses).
However, what’s really broken down earlier resistance is our old friend complexity. To get a multi-million gate design out on time today probably requires not just one or two of the options available, but a combination of all four. The trick is in knowing which one to use, when to use it, and what part of the project it will best serve.
And that is where Synopsys may think it has (and indeed may have) stolen a march on its rivals by buying EVE and its ZeBu product line. The deal doesn’t just give it the ability to market a very tightly integrated soup-to-nuts flow, it can also arguably construct this flow in ways that are more adaptable for achieving the right balance. The deal could prove to be as much about design services as black boxes.
That said, everybody knows the problem the clients face, so no-one should expect either Mentor Graphics or Cadence Design Systems to sit idly by. Indeed, could we be on the verge of seeing some further acquisitions in areas where they are not direct players today? For sure, neither company will answer that (they’d be mad to), but the question stands. Certainly, they’ll be looking to add still more finesse to answering, “What do I do and when?”