And Cadence should buy Mentor--and for good reason: the troubled EDA company needs to beef up its design-for-manufacturing (DFM) offerings, said Raj Seth, an analyst with SG Cowen Securities Inc. (New York), in a new report.

As reported, Cadence made a bid to acquire Mentor. But Cadence's $1.6 billion offer for rival Mentor Graphics is seen as too low by Mentor's management, which also said it is concerned regulators might scuttle the transaction for antitrust reasons.

Analysts see the handwriting on the wall for Mentor, despite its resistance in the surprising Cadence bid. ''In our view-- especially with a Mentor board meeting around the corner--it will be difficult for Mentor management to fend off Cadence, especially if Cadence has some capacity to increase the bid,'' Seth said in the report.

''We think this deal could get done,'' Seth said. ''Expect Mentor to hold out for more money, and believe that Cadence could pay $20 per share and still make it work.''

The analyst also sees no White Knight in the making. ''A counter offer from Synopsys seems unlikely, but remains possible. That said, we believe Synopsys is in better fundamental shape than Cadence,'' he said.

In some ways, a Cadence-Mentor merge makes sense. ''Mentor has a large direct sales and support channel,'' he said. ''Taking a large competitor out has positive implications for industry pricing which benefits everyone.''

But here's the real key for Cadence: ''The jewel in the Mentor portfolio is its Caliber physical verification offering and its strong position in the design for manufacturing market,'' he said. ''Cadence is weak here. Mentor also brings some more focused solutions in automotive and other verticals which are appealing to Cadence.''

Recently, Mentor laid out its IC implementation strategy to help chip makers move to the smaller process nodes. A first step in this strategy was the merging of several offerings into a single Design-to-Silicon division. This includes its prized Calibre physical verification and design-for-manufacturing (DFM) platform.

The Calibre DFM solution includes the Calibre LFD (Litho-Friendly Design) capability which is integrated with new Calibre LVS ADP (Advanced Device Parameter) features. The combination of the Calibre LFD and LVS capabilities provides a complete and integrated solution that improves SPICE simulations by providing more accurate results that better reflect actual silicon performance, according to Mentor.

Another piece to the DFM puzzle is Mentor's Olympus-SoC place and route product. It is said to accelerate signal integrity closure. The multi-corner, multi-mode (MCMM) capability of Olympus' Static Timing Analysis (STA) engine concurrently computes delay shift and glitch for any number of mode/corner scenarios in a single pass.

There is, however, some overlap between Cadence and Mentor in verification and emulation. ''There is substantive overlap in areas like PCB design. We think Cadence would have to divest this,'' he said. ''Integration, especially with very different corporate cultures wouldn't be trivial.''

Others have another viewpoint. Cadence is under pressure and may lose its top spot in the EDA sector, but its proposed $1.6 billion merger with Mentor Graphics is "a really bad idea," said a veteran EDA analyst.

The analyst, Gary Smith, principal of Gary Smith EDA (Santa Clara, Calif.), in a recent interview said that Cadence has basically lost three of its four ''franchises'' in recent times, including ASIC layout, design rule checking (DRC) and Verilog.

In those markets, Cadence has lost share. Its last ''franchise'' in EDA--analog and custom design--is under pressure from various forces, Smith said.