Internal EDA tools can fill some important gaps in the hardware design ecosystem, according to a recent email survey sent to members of the Cadence Community. In the survey, 256 respondents first indicated whether or not they use internal CAD tools. Those who said they did – about one-third of the total – then went on to answer questions about whether usage is changing or holding steady, how much of the design flow is comprised of internal tools, why they use internal tools, for what tasks they’re used, and what the primary drawbacks are. The survey followed a discussion about internal EDA tools in a previous Industry Insights blog.
Key takeaways from the survey include the following:
A minority – but a significant minority – of designers use internal EDA tools. Internal tool use seems to be holding steady, as opposed to notably increasing or decreasing.
- In most cases, internal tools comprise less than 25% of the design/verification flow.
- The main reasons for using internal tools are to provide capabilities not available commercially, and to serve specialized applications. Cost savings is a lesser reason.
- Internal tools are as likely to be developed directly by design/verification teams as by internal CAD departments.
- Internal tools are used for a broad range of tasks, with no one task clearly dominating. Functional verification topped the list in our survey.
- Internal tools have some drawbacks. The biggest one is that the people who designed them might no longer be around.
The first question shows widespread but not majority usage of internal tools. In contrast, a Gartner Dataquest study in 2005 showed that 27% of respondents use internal EDA tools, while an EE Times 2005 survey put that number at 56%.
Answers to the following question indicate that internal tool usage is not dramatically increasing or decreasing:
Below, we see that in nearly three-fourths of design flows, internal tool usage is less than 25%. Still, there are some flows comprised mostly of internal tools.
Internal tools are most often used to provide capabilities not available commercially, or to serve company-specific applications. Some of the “other” options cited for question 4 include “can customize tools,” “superb support,” “open source solutions are better,” and “job security for internal groups.”
The next question may come as a surprise to those who assume that internal tools are developed by internal CAD groups. Only in some cases is that true.
Question six shows that internal EDA tools are used for a broad range of applications – there’s no one capability that really stands out. In contrast, in the 2005 EE Times study mentioned above, “system-level design” was clearly the leading applications area for internal tools. That may have changed with the maturation of commercial ESL tools.
“Other” options cited include floorplanning, design for test (DFT), signal integrity, project management, reliability, mask data verification, IC layout, and design translation.
We also asked about limitations of internal EDA tools. The top two responses below have to do with maintenance and support. Some respondents also noted that internal tool development is costly – an interesting contrast to those who said they use internal tools because they’re “less expensive” than commercial tools.
There were also some verbatim comments on the overall survey topic. A sampling is shown below.
“Internal tools allow us to adapt to problems in a timely manner. Developers are responsive and customer driven. Problems can be identified and fixed in days rather than months. Cost 1/5 of external tools. Only limitation is resources for internal teams.”
“We have a few [internal] EDA tools that have been around for a long time and still provide value for an incremental cost of development. We probably would not make the decision to develop these as new tools if we were making the decision today. Most of the new tools we develop are for very specific design tasks that we haven’t found commercial alternatives for.”
“We’ve always developed tools to work around shortcomings of commercial tools, or when we were unwilling to tie ourselves to a single vendor implementation. Most internally developed tools become commercially available eventually, and once they become standardized, we often migrate.”
The last two comments suggest that internal tool development is largely undertaken when commercial capabilities are lacking, and that most design teams will generally prefer commercial tools. There is no convincing evidence of a shift away from commercial tools. But internal tools will continue to be part of the landscape, and they may help pioneer leading-edge capabilities that will eventually find their way to the marketplace.