Plunify: FPGA Design Closure

The number of EDA startups has plummeted around the globe, and nowhere is this more evident than in Singapore. In fact, there is exactly one EDA startup in that country—Plunify—and even that isn’t so new.

Plunify started life in 2009 as a cloud-based startup, whose mission was to provide public cloud compute services to companies developing FPGAs. While this approach seemed like an obvious way to provide incremental revenue, both to startups and big EDA companies, it hasn’t worked for anyone. Chipmakers don’t like to use public infrastructure—or even private infrastructure outside of their own data centers—to analyze, verify or debug their code.

So last June, with backing from the Singapore government, Lanza Ventures, the company did a soft relaunch focused on design closure.

“About 80% of the FPGA challenge is getting it to meet timing,” said , Plunify’s founder and CEO. “People spend a lot of time changing the source code or trying brute force approaches. Our tool helps them to do this without changing the source code.”

So far, he said there has been strong interest from telecommunications companies for wireless bases stations, oscilloscope makers for industrial control, automotive companies and military and aerospace companies.

“There’s also a big role in data centers, where you can offload server communications, and we’re getting strong interest from high-speed financial trading houses,” said Ng. “For them it’s less about getting the design to meet timing and more about how much faster they can get their designs.”

Plunify’s technology uses data analysis, probability and machine learning to achieve optimization. “We’ve only scratched the surface for what this technology can do,” said Ng. “A design team looks at a problem, and then they put all hands on deck to find a bug. With this technology you can start it running before you leave the office and when you come back in the morning you have an answer. There is no other way you can check the hundreds of thousands of combinations, and the answer is not to hire more engineers.”