September 12, 2005
Who Is Using ESL and Why?
Please note that contributed articles, blog entries, and comments posted on EDACafe.com are the views and opinion of the author and do not necessarily represent the views and opinions of the management and staff of Internet Business Systems and its subsidiary web-sites.
| by Jack Horgan - Contributing Editor
Posted anew every four weeks or so, the EDA WEEKLY delivers to its readers information concerning the latest happenings in the EDA industry, covering vendors, products, finances and new developments. Frequently, feature articles on selected public or private EDA companies are presented. Brought to you by EDACafe.com. If we miss a story or subject that you feel deserves to be included, or you just want to suggest a future topic, please contact us! Questions? Feedback? Click here. Thank you!
Who is using ESL and why?
On August 31 Celoxica announced the results of its second survey on ESL usage conducted between mid May and mid July 2005. In all 723 designers responded to more than 30 questions about current and future design activity.
I had an opportunity to discuss the survey and Celoxica with Jeff Jussel, VP Marketing and Americas General Manager. Jeff joined Celoxica in 2002 as VP of Marketing. Prior to Celoxica, Jussel worked at Mentor Graphics, serving the last two years as head of worldwide marketing for the professional services division. He has over 18 years experience in electronic design serving in engineering, marketing and management roles. Celoxica provides C-based behavioral design and synthesis solutions along with boards, IP and services to companies developing semiconductor products.
Who was surveyed and how were they selected?
There were several different ways of getting people. Basically we had Electronics Weekly UK promote it. We had it on our website for the entire period (June-July). We also sent it off to the DAC email list and the Techonline email list. There was a broad targeting of people both with software and hardware backgrounds.
What was your motivation of the survey?
We've noticed through our business and going through the different projects that we do that the adoption of ESL is coming from across the board not necessarily directly from the hardware people but heavily from embedded software that are not traditional ASIC designers. This is the second survey we've done to track that trend and to see if the industry at large has the same experience that we have.
Who participated in the survey?
We took it to the people we mentioned. There were over 700 responses. From that, not surprisingly, 50% or better were hardware people. However, 40% of them were not. Software people, verification engineers, system engineers and algorithm designers were high on the list as well. So even though there was a preponderance of hardware people - not too surprising, since we relied on the DAC list and DAC people are going to be costly hardware engineers. We still got a big response from the other group.
What was the geographic split of responders?
We were split pretty much half and half, Europe and the US. We had 15% from APAC. Based on the people that we sent it to and the fact that it is an English language survey, it is not surprising that we didn't get much more from Asia and Japan but still enough to be representative.
Describe the results of the survey?
We found out that people are mostly designing things that are embedded systems, FPGAs, on-chip processors or programmable systems. Few of them are solely ASIC designers as we would have expected maybe 10 years ago. Most people are embedded system designers who are looking at things that are more programmable, more software oriented. That corresponds to what we expected as well. If you look at the industry overall, we a seeing a move towards embedded systems including IP reuse, the use of programmable logic and the use of more processors if you can get away with performance using a processor rather than custom ASIC. I think that corresponds to the rising cost and complexity of going
into ASIC designs where there are few applications where you can justify the risk of going to custom ASIC. If you can get away with using programmable logic, another processor or existing IP you will. We are starting to see that more and more. The percentage in the ASIC design space is going down. Certainly other surveys that we have seen have corresponded to that as well. There are less than 2,000 design starts on the ASIC side while non-ASIC or FPGA design starts are orders of magnitude higher than that. We've seen numbers like 80,000 per year and growing. That doesn't include all the processor and embedded processor types of designs that we and others are not even capturing with
As far as those responders that we have the majority are planning to increase their use of ESL design. The one's that are planning to increase their use of ESL design are not necessarily hardware designers. 52% of hardware designers agree that they will increase their usage but very heavily skewed towards what we call non-traditional hardware designers which are the other developers, the system engineers, the algorithm developers. 50% of algorithm developers and system engineers are going to be using ESL and as well as 50% of software engineers. This helps explain why we are seeing more growth of companies like MathWorks who consider themselves to be ESL companies because it's their
traditional market. The algorithm designers and the researchers are moving more into implementation. They are taking advantage of programmable logic in order to develop their algorithms. They are the one's who are going to be suing ESL even more so than hardware designers. It's a good news for EDA companies that are in ESL because traditionally the EDA market had a few thousand hardware engineers that you can sell to and it adds up to maybe $4 billion per year marketplace. But if you add in all the algorithm developers, scientists, system engineers and software people you've added 10x the number of people that can be a target market for ESL. Based on that, we have the opportunity for
EDA to grow well beyond its current market. That agrees with what Gartner is reporting as well as far their expectation of ESL.
What are designers designing?
If you look at ESL from the hardware point of view you might think maybe this is all about productivity. We are actually seeing ESL being used for access where algorithm developers, system modelers who wouldn't necessarily have had access to hardware models before now have the ability to target things from a software based model. ESL gives those people access to using custom electronics where they wouldn't have been able to do that before. They would have been stuck with a processor. Now they can split that between a processor and maybe customizable programmable logic or a structured ASIC. Now that they have that capability by using ESL techniques. Those are the things we are seeing
We are also encouraged by the fact that more people are actually experienced in this area. When we did this last year fewer people were actually in evaluation or pilot projection to production phase. We had a lot more people in the intending to use phase. We are seeing that greater than one half the people are in the evaluation and production phase actually using the product. We are seeing adoption across the board. That adoption is having an impact on ESL languages which is why you are seeing ESL languages pickup tremendously. System C did very well in the survey not too surprisingly because the audience was 60% hardware engineers. The one thing that hardware and software people
could definitely agree on was that System C is the language. Some of the other languages we think will pick up even more around algorithm developers, system engineers but System C is definitely popular among both. Not a surprising finding. But something that might be surprising to hardware people is the forecast that the use of HDL will decline. System Verilog was also forecast for only small increased usage. We're seeing that because System Verilog is basically in its capabilities another RTL, a hardware description language. ESL languages by definition have to be able to program both hardware and software. System Verilog doesn't fit that bill. We don't have any compiler that goes
to an ARM or whatever, not from the System Verilog language. It's really a kind of misnomer. We are seeing that from the survey as well. Verilog designers are adopting it and there is a little bit of increased usage but basically it's just another HDL.
You can find the full EDACafe event calendar here
To read more news, click here
-- Jack Horgan, EDACafe.com Contributing Editor.
Be the first to review this article