Tuesday, March 24, 2009

EDA Industry Update

Sramana Mitra posts her latest financial analysis of the leading EDA vendors at EDA Industry Update.

Some reactions to her points:

  • Synopsys is king of the hill. Yes, obviously.
  • Cadence should acquire Magma. I could see this happening. That's the kind of thing Cadence has done in the past, and Magma is getting small.
  • Bring back Joe Costello. It would certainly energize the industry, but I doubt he'd want to, or that he could fix things alone.


Harry Gries said...

Sramana says "Rajeev has reconciled with the fact that he needs to get acquired soon, and that Cadence would be his preferred new home for Magma".

5 years ago Magma was the beautiful young princess and any prince would have gladly taken her to be his bride and eventual queen of the kingdom. But Magma wanted the entire kingdom for herself, so she shunned all the suitors and said "no, I do not need a prince. I will rule the kingdom on my own."

Now, Magma is much older and less attractive. What prince really wants her now? Especially since all the princes are already married (Cadence Encounter, Synopsys IC Compiler, Mentor Olympus).

I don't see any of the "Big 3" buying Magma except to get the technology in order to merge into their products somehow. If you forced me to pick one, I think it would be Mentor who does not have ASIC synthesis, has a promising but 0 market share P&R solution, and has no mixed-signal design solution. At least then Magma would fill some holes.

Nick said...

Hi John,

ps: please dont consider this to be a cadence hate speech!

Rajeev did manage to made cadence a scape goat when he sold them Ambit synthesis (for 50 million bucks). I am pretty sure he is capable of pulling this off once again.

Cadence went and bought Get-to-chip for another 150 million. For 200 million bucks, they had a return on investment of 10 million and ended up with about 2% of the total ASIC synthesis market.

Whoa! well done!. They look like Lehmann brothers of the CAD industry to me!

I have to say that, after 2 product acquisitions and 15 years of Cadence R&D effort, they succeeded in building an incremental synthesis tool (i.e synth -incr) which worked on DC netlist's and produced netlist's which looked fairly optimized and better (i.e with relaxed wireload models and until they hit physical synthesis realm).

Looks like Cadence has a nice/ relaxing work atmosphere in their logic synthesis team and sales teams which is not really results driven.

In comparison, I consider magma to be a fairly competent tool.

Doesnt matter if it started off with Lukas Van Ginneken taking his ideas for a spin around the world (i.e->IBM->SNPS->LAVA).

Cadence can surely acquire magma and then have a total of 2% of the total physical implementation market. They proved it once with logic synthesis. Now they will try a hand at physical implementation.

They seem to be are fairly good at mathematical induction.

I would rather prefer to see a smaller magma come out stronger through this crisis with better products than ever before and take on synopsys.

Magma can survive!.The need to shed all the extra fat that they have put on around their waist since 2003-2004, scrap underperforming products and focus on their core competence (which is physical implementation).

If Magma gets acquired by cadence, Synopsys will truly be the ASIC "king of the hill".