August 21, 2007 – The acquisition of DFM darling ClearShape shows Cadence is finally getting serious about DFM, according to analyst and EDA guru Gary Smith. Although there are still some areas to address, he says, the DFM market now seems to be a two-horse race until competitors answer back.
“This really puts Cadence now into the DFM business,” said analyst Gary Smith, of Gary Smith EDA. Their recent purchase of Invarium “was a good buy, technically,” but adding Clearshape, “the darling of the DFM market,” changes the game.
The need for DFM is magnified by the shift from 65nm to 45nm manufacturing processes, which also means moving from rule-based to model-based DFM, Smith explained. Rules become too complex and contradictory, necessitating a new model of the process, running simulations to see what will happen. The key to that is also making those models secure — “if you can break the encryption, you have [the chip manufacturer’s] process information — if you were so inclined,” Smith said.
Another DFM vendor, Brion, gets around this problem by giving its technology to customers wrapped inside computing hardware — something that’s appealed to IDM customers, Smith explained. Clearshape, he said, seems to have figured out how to build a secure model-based process, that’s also proven accurate for 90nm and 65nm, and again at 45nm. “For foundries and fabless companies, ClearShape has made a lot of converts,” he said.
The deal not only represents “the new Cadence,” a company that’s trying to reinvent itself in the face of strong competition, but also the evolution of a DFM company that sees its future as part of a larger ecosystem where design and manufacturing coexist, Smith said. Plus, “Cadence has a lot of clients that use primarily Cadence tools, [and who might be] tempted to now swing back to Cadence” — including foundry giant TSMC, Smith noted.
ClearShape, meanwhile, realized the next phase of its evolution wasn’t going to come via IPO (the DFM market “will be settled in the next two years,” Smith noted) or through the risky path of M&A to achieve critical mass like Magma and Avante had done. “They saw that Cadence has really turned around in the last six months with significant R&D efforts,” Smith said.
Another advantage to ClearShape combination is it’s the true “design-for-yield” tool, Smith noted. There’s still uncertainty whether and when designers will actually use DFY, Smith said, noting that “design-for-test took seven or eight years before designers would use it — and they were forced into it.”
There’s still an area in Cadence’s platform that needs to be addressed, according to Smith. Cadence’s Open Access database has “never really been involved in any post-GDSII work,” Smith said, and a new version of Open Access will require an entire rewrite to work with the parallel architectures needed for 65-45nm and beyond. But until design engineers prove they’re actually using the DFM tools, that won’t be a priority yet, he said.
So what could happen next in the DFM space? With the industry’s “darling” now locked up by Cadence, “everyone will look at Blaze” as the next prize DFM firm still out there, Smith said. What happens next really is up to the other big guns: Mentor, Synopsys, and maybe Magma. “Mentor has their own model — they won’t talk about it, but they have it,” Smith said. “Synopsys, I don’t know what they’ll do. For the forseeable future, in the modeling business, it’s Cadence vs. Mentor.”