Since the beginning of Synflow (back in 2012, before we even incorporated the company), we have been trying to get people to use our software, and give us feedback about what worked, what didn't, what they expected, etc. At the same time, we were also meeting prospects face to face to explain our products so they would eventually buy them.
Some of the prospects we met were designers working for French R&D subsidiaries of the top 10 semiconductors companies. Despite working for semiconductor companies, these designers were open to innovation, and were well aware of the increasingly difficulty to meet deadlines with existing hardware processes and tools. They were still using Verilog, System Verilog, and SystemC simply because of the lack of efficient alternatives.
They expressed their feelings and frustrations with High Level Synthesis: unpredictable, steep learning curve, inappropriate input languages (synthesizable C/SystemC), etc. When they realized that Synflow was not doing another HLS tool, but rather a new framework for hardware design and verification, they showed interest and provided us with valuable feedback.
Understanding is the reflection of creating - Auguste Villiers de l'Isle-Adam
Unfortunately, a lot of hardware engineers are not that open to innovation, or at least not when it's coming from a start-up. Something I often hear is "I prefer not to change our flow because it works"; and this is when I hear back from people, but more often than not people do not even dare answering emails. On the one hand, this is something I can understand: the fear of change, a "zero-risk" strategy, a disappointment with HLS technology. On the other hand, when you discover a new framework that can make you more productive, make your daily work more fun, and when you can try it for free during two weeks, is it really not worth trying?
Of course it is worth trying! And this is exactky what our users from the software world are doing. So why are these guys using our software, and not hardware designers? Culture. Software engineers know that you don't need to be an old company to provide quality products, and that a jump in productivity often comes from new technologies, languages, and frameworks. This is what Cx gives them: being able to start designing hardware with Synflow in a couple of hours when it normally takes months to learn traditional HDL flows or HLS.
On the opposite, many hardware engineers look with mistrust at new technologies (and start-ups), and wait for traditional companies to innovate. Xilinx Vivado HLS is the perfect example. When Xilinx bought AutoESL and rebranded the product, it became a success. The same product, but provided by Xilinx, raised the general trust in reliability. This difference in culture and tools is well described by John Blyler in this article.
What does the future hold?
When the time to market makes the difference between success and failure, staying behind worrying about whether we should adopt a new technology or not can be a deadly mistake. In recent years, new languages and frameworks were created to develop software apps and websites. This has driven the emergence of a wide variety of start-ups, new companies, and success stories, and at the same time it is driving slow-moving, change-adverse companies down.
I can hardly wait to see if the majority of hardware engineers will keep walking the same old path or finally jump on the innovation bandwagon.