Is the Asic over?
“Can anything save Asic?” was the question asked at a panel session at the Globalpress Summit meeting in Monterey.
Dennis Kish, senior v-p at Actel, pointed out that the Asic market was worth $23bn last year and was forecast to grow faster than the overall semiconductor market between 2002 and 2010, while Hugh Durdan, v-p at eSilicon pointed out that, if you added in the $60bn value of the ASSP market, then a very large market is being discussed.
Steve Carlson, v-p at Cadence, reckoned it is changing, not dying. “I was at LSI Logic. The LSI Logic model is dying a slow and painful death but a number of new models are being experimented with.”
Elie Massabki, v-p at ChipX said: “It’s a $20bn market which is not going away. The market is shared between full custom, standard cell, gate arrays, structured Asics. Some are declining, others are growing.”
Durdan of eSilicon asked: “What is an Asic? If you take the definition of a custom chip done by a system OEM to add value and differentiation to a system, then that is in decline.”
However, he did not see that as favouring FPGA. “Only ten per cent of the cost of developing a chip is putting it into silicon. The rest goes into defining architecture, software, verification etc,” he said.
Naveed Sherwani, CEO of Open-Silicon, said: “The real question to ask is: Why did Asic survive? If you look at the ridiculous cost, the extreme unpredictability, and the extreme unreliability, Asic vendors did everything they could to kill the business. The numbers are shameful, really shameful. The Asic vendors have survived because they were needed. You can never do in software what you can do in hardware. I have been a designer all my life. You can do magic with transistors.”