|
Viewing 1 to (4 Total) Moshe, Krishna, and Kevin talk about FPGAs |
Total Posts: 84
Joined: Apr 2009
|
I had a great chat with Moshe Gavrielov and Krishna Rangayasee of Xilinx about the future of FPGAs, the decline of ASICs and ASSPs, and Xilinx's role in the whole thing.
The article is here
What do you think?
Posted on 2010-01-06 02:02:39 at 2010-01-06 02:02:39
|
|
Total Posts: 21
Joined: Dec 2009
|
The interview
Interesting interview with two very important players in the FPGA game.
I especially found the last question/answer telling. I don't think they answered that very effectively, to be honest. They basically said that there wasn't enought money for those FPGA startups, but there are a few and some of them have VC money. I'd sure like to see an FPGA Journal article listing the new ones and what their differentiators are.
I've got to wonder, though, if these two really understand who drives the decisions about what products or architectures are used in a design.
Of course, now I've got to wonder if I understand who drives the decisions about what products or architectures are used in a design in most companies. 
They pointed out that they are getting more respect, that is, getting more ear time with CEOs and the like. Does that matter all that much? Few things annoy an engineer more than having the vendor go to the higher level managers and convince them the product is the right one -- even if it is. It results in too many Dilbert moments, too much time spent explaining why the vendor is wrong in this case or that. I think it's still true that you need to sway the engineers. Let them sway their managers.
They need to convince engineers that their FPGA is the first choice among many, rather than the alternative choice.
Cliff
Posted on 2010-01-06 11:28:48 at 2010-01-06 11:28:48
|
Total Posts: 1
Joined: Jan 2010
|
While accurate, there is one thi
While accurate, there is one thing that is being ignored regarding investment into new programmable logic technology - the customers. If you spend time with the leading programmable logic users in the world - the "Ciscos" ("Huaweis"?) as well as consumer product vendors such as Samsung and LG - you will hear that they are frustrated with the difficulty designing with today's FPGAs from A and X. Timing closure for high-performance designs can take as long in a FPGA as in an ASIC. At the other extreme, customers want to innovate in specific areas and leave the rest to the vendor (ASSP model), but they can't get that level of support. FPGAs have done a good job straddling the middle of these extremes, but now they are vulnerable to solutions that will serve these extremes better as measured by customer frustration. Market opportunity dictates investment, not VCs. Having said that, I hope A and X see the opportunity themselves and invest!
Posted on 2010-01-06 13:10:16 at 2010-01-06 13:10:16
|
Total Posts: 14
Joined: Nov 2009
|
middle ground
I think the large product vendors are the only companies that actually benefit from ASICs from a cost-savings perspective because they buy in such large volumes. That's why Xilinx is structuring their business model to smaller companies with less VC money. So in a sense, they are not marketing the extremes but rather the broad middle ground - at least that's the way I understood it.
What I think is interesting is they spoke about how service is becoming just as important as hardware. I think this is a natural shift because smaller companies don't have the resources or time to develop all the IP needed for complex designs. With more advanced tool-sets and IP, more demand for service will inherently come along with it.
Posted on 2010-02-23 13:50:03 at 2010-02-23 13:50:03
|