You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
FPGA is going to be around. Xilinx is the market leader. I think Intel buying Altera means that they have limited competition in most of their markets since I think Intel will focus on the datacenter. They are also very strong in the software side of things. See my blog last week. And if you are worried about the short term (their revenue has been declining) Goldman just upgraded them and thinks this will turn around.
Yes I read about Xilinx being more of a software company. Well, I'm a circuit designer so I am more worried about the semiconductor side of things. Intel buying Altera can go both ways for Xilinx. Right now, some customers are switching to Xilinx. Also, for volume productions, FPGA aren't a good solution so FPGAs can never be used for mobile or things like IoT etc.
Automotive, healthcare etc? Yes. I wonder how the growth chart of FPGAs vs ASICs will look like by 2019.
Yes I read about Xilinx being more of a software company. Well, I'm a circuit designer so I am more worried about the semiconductor side of things. Intel buying Altera can go both ways for Xilinx. Right now, some customers are switching to Xilinx. Also, for volume productions, FPGA aren't a good solution so FPGAs can never be used for mobile or things like IoT etc.
Automotive, healthcare etc? Yes. I wonder how the growth chart of FPGAs vs ASICs will look like by 2019.
ASICs will have a better growth rate but there will always be more FPGA design starts. Mobile, wearables, and IoT will drive ASIC design starts moving forward. My advice would be to go ASIC. If you are going to go FPGA, Xilinx is the place to be. Xilinx and TSMC have only been working together since 28nm so they are just now hitting their stride together. Xilinx beat Altera at 28nm by 3 months. Xilinx beat Altera to 20nm by 1+ year. Xilinx will beat Altera to FinFET at 16nm and 10nm. My opinion.
I worked at Altera for a short period of time when they were trying to get Stratix V up and running.
I probably cannot say what happened there.
Later on, I had an opportunity to ask several Altera employees on how the Stratix V launch went.
Pretty much all of them pointed out to variety of problems in execution (i.e., design software stability, SERDES issues, etc.) during the period (I had 3 sources.).
As far as I can tell, Altera has not announced the tape out of Stratix 10.
They are already 2Q behind Xilinx, and I read that they are now looking for 2H15 tape out.
Since there are many things that cannot be tested until the silicon is ready, I will not be surprised if the MP of Stratix 10 will get pushed into 2017.
It is not looking good for Altera, and maybe this is why Intel bought them.
In case of Xilinx, my source (a Xilinx employee in the SERDES department) told me that it took 3 months to get the first functioning 7 series FPGA (Kintex-7), but it only took 3 days to get the first UltraScale FPGA working.
