Between The Lines

Hot on the heels of posting that Cadence was rumored to be buying CommandCAD (an acquisition I later confirmed), John Cooley has just posted another EDA M&A rumor. This latest one is that Mentor Graphics has acquired Sierra Design Automation. I have inquiries out to Mentor and contacts, so stay tuned...

A few years ago, Cooley started bagging on Mentor because while Mentor offers just about every other type of EDA tool (IC design, IC verification, emulation, DRC/LVS; PCB design, verification and layout; FPGA synthesis; and even wire harness technology), it does not offer a IC physical design suite. Cooley dubbed Mentor “the donut company,” because of the hole in the middle of its tool line up. Indeed, it is a bit strange that Mentor plays in just about every other segment of EDA (and even plays in some that aren’t EDA like embedded software and silicon IP), but doesn’t play in IC physical design, which historically has been one, if not the most lucrative segments in EDA.

If memory serves, Mentor didn’t like being called the donut company and grumbled a bit to Cooley about the moniker. Since then Mentor execs all adopted a company line to inquiries about the hole in its lineup by saying “we only play in areas where we can hold dominant market share.” That’s certainly the case with DRC/LVS, the VHDL half of HDL simulation, wire harness tools, and maybe a few others but it also plays in areas, where at last word, it doesn’t own dominant share such as Verilog simulation, emulation, IP, PCB, analog simulation and layout, embedded software…

But quietly behind the scenes, Mentor HAS in tried to enter IC physical design. Mentor was very close to acquiring Avanti shortly after that company’s legal troubles ended. Mentor pulled out of the acquisition when Avanti tried to turn it into a bidding war with Synopsys. That acquisition would have given Mentor dominant share in place and route (at least with power users of the time) and in a few others.

What’s interesting today is that the IC physical design landscape has changed. Mentor’s built a powerhouse out of its Calibre lineup and to its credit has been very diligently changing with the DFM times to not just maintain but grow Calibre’s share. Mentor acquired OPC technology very early in the game and ever since has added a bunch of incremental upgrades to the tool to keep it going strong over the last half dozen years. But I don’t think it’s a stretch to say that all the neat DFM stuff we see today as point tools will need to be integrated and run transparently in the next generation of physical synthesis, floorplanning, place and route tools.

My guess is that the next big winner IC physical design will implement all these OPC and litho correct features during routing, not after GDSII and the features will be under the hood and transparent to users. My guess is that to make this happen, the router and DRC/LVS tools will need to be finely correlated. And there’s the rub for Mentor and Calibre. If you have a DRC/LVS tool but don’t have a P&R tool, you’re at a bit of a disadvantage. Cadence, Synopsys and Magma have both types of tools (physical design and physical verification) and all three share a fairly equal chunk of IC place and route market. Meanwhile, Mentor clearly has the dominant share in DRC/LVS. So if Mentor were able to field a merely viable P&R tool, it would seemingly have an advantage. If it were able to adapt to the changing requirements of silicon and needs of designers like it has with Calibre, look out, you have a new contender for the P&R title…well, maybe.

Does Sierra have a killer physical design suite that can turn into a marketshare leader? With Mentor’s help… maybe.

Last July (10 months ago, Cooley), I suggested Mentor would need to pair up with a P&R company and to watch Sierra.

I don’t know if the Mentor/Sierra rumor is true or not but if it is, it will be fun to see what Mentor can do without a hole in its lineup.