Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Xilinx (XLNX) -- Ignore unavailable to you. Want to Upgrade?

To: SemiBull who wrote (2888)	11/18/2002 7:48:54 PM
From: SemiBull	Read Replies (1) \| Respond to of 3291

The Interview: Wim Roelandts, Xilinx CEO, sticks to FPGA basics By Peter Clarke, Semiconductor Business News November 18, 2002 URL: siliconstrategies.com MUNICH -- Although in 2001 FPGA vendor Xilinx was briefly caught in the down-draught of the communications bust it has generally managed to turn in better financial numbers than most of its ASIC rivals. At last week's Electronica exhibition SBN caught up with Wim Roelandts, the company's president and chief executive officer, and asked for a quick run down on why, in difficult market conditions, Roelandts thinks Xilinx is in good shape, and what needs to be done to extend the company's position SBN: Over recent years Xilinx has been quite successful at breaking into some consumer electronics opportunities. Do you have a view of what sort of holiday selling season there is going to be? Roelandts: Well of course most of the manufacturing for holiday goods has already taken place, the question is whether there will be replacement sales for more goods. We think the holiday season has got underway but it will be down a little bit [on previous years]. SBN: How are markets doing regionally? Roelandts: Well it is clear that the U.S. is weak, Europe is slow, but Asia-Pacific is very strong. Overall there is a lot of uncertainty about. But economic instability is good for us. People are fearful of [starting] ASIC designs, so they turn to FPGAs. SBN: How many design starts does Xilinx see a year? Roelandts: We call them sockets because some designs could include several FPGAs on a board. We win more than 20,000 design sockets a year although only about 12% or 14% go to high volume. We're doing pretty well. Since our Virtex-II range of FPGAs was launched in 2001 it's done about $160-million of revenue. And most of that is still prototyping business. SBN: With the advent of Virtex-II and the offer of hard diffused cores within Xilinx products, the company's identity seems to have become harder to pin down. There are systems considerations, multiple alliances, with IBM on PowerPC and so on. How would you characterize Xilinx now? Roelandts: We are 100% about FPGAs. But we do have a tremendous arsenal of IP [intellectual property]. We've got a couple of hundred of pieces of IP that have been demonstrated to work. We also offer design services. We also offer things like development boards for VxWorks [embedded operating system] from WindRiver, and embedded Linux from Montavista. We show users what can be done and where they can go get it. SBN: Does Xilinx need to move up to address high-level language based design? Will you buy Celoxica? [developer of the Handel-C high-level design language] Roelandts: Probably not [buy Celoxica]. The EDA software we do is tightly connected to the place and route. Deep-sub-micron silicon is getting very complex, but in FPGAs we take care of all that for the developers. SBN: But surely there is a need to cut design times, to offer new software to handle the complexity that is becoming available on FPGAs? Roelandts: It doesn't seem to be a high priority with users. We do support HLLs such as C but we don't do it ourselves. We leave it up to Celoxica and others. There is no need to put our money there. After all VHDL is a high-level language. We also work with Matlab and that works well. SBN: Do you have any new major FPGA families in development? Roelandts: Every new generation of FPGA contains new features. They are complex systems that evolve. Coping with Moore's Law plays its part so we always look to optimize routing, logic blocks, clock management and other functions. But there's no new architecture that is much better coming along. For example, we can use nine-layers of metal, we use copper, more than the ASIC developers. We use UMC and IBM for volume. Some 31% of our output is made on 300-mm wafers.