With semiconductor chip maker Intel Corporation introducing the world's first 20-nanometer (nm) 128-gibabit (GB) die in December 2011, two of its largest global competitors – Xilinx and Advanced Micro Devices (AMD) – also from the US, are gearing up to join the league. And, the hub for the evaluation-to-execution activities will be their research and development (R&D) centres based out of Hyderabad.
“Our team in India (Hyderabad) is starting to look at the next-generation 20nm process products. This is something that you will see from Hyderabad in the next two years. We are in the planning stage at this point and cannot go into anymore details. All I can say is that the team is exploring areas,” said Suresh Menon, vice president (product and programmable platforms development), Xilinx.
Xilinx, a player in the digital programmable logic devices market, has its largest R&D centre outside of its US headquarters in Hyderabad, with 400 engineers working for its core field programmable gate array (FPGA) platforms. As part of its 7Series’ 28nm product portfolio, the over $2.4-billion company has launched globally its Virtex 2000T, which it claims to be the world's highest capacity FPGA. It has started sampling since last quarter.
“At every node we were one of the first companies to take out the products. The cost at each node is going up by 50 per cent or more, which not many companies can afford to. So, progammable logics have an immense opportunity to sustain these costs. In 28nm we are either number one or number two globally and we expect to be in that league in the case of 20nm too,” Menon said.
According to Manoher Bommena, site head (Hyderabad), AMD, a global provider of processing solutions in computing, graphics and consumer electronics markets, the company's 650-strong R&D team in Hyderabad is working on 28nm and was looking into doing some test chips for evaluating the 20nm technology-readiness for its products.
“Now, people who have more research-oriented mindsets are looking at future technology and their viabilities to our products. Lots of foundries are already working on 20nm technologies. Once they give some stability to their process and requirements ... that’s when AMD will come up with its products,” he added.
Elucidating the benefits that the 20nm process technology brings to the table, Bommena said the smaller the geometry the lower the power. “With more and more components being pushed into one board, the real estate on the board comes down, and that requires less power. Typically, we can expect 10-15 scale down in power utilisation with 20nm,” he said.
