David Dill

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

Again, NSF is responsible for thousands and thousands of innovations that have made the modern world. Murphi made an important difference, but it's a tiny fraction of the work funded by NSF. For this reason, NSF has been enthusiastically supported by Democrats and Republican alike.

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

I think the first Murphi paper was published in 1991. Recently, I was contacted by another former student that Murphi was being used at a very important computer company <redacted>, and that it was still regarded as the best tool in industry for debugging cache coherence protocols.

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

In the early years, I'd hear from people that it was being used routinely at their company and there were local experts who used it full time to debug their protocols. At some point, it dawned on me that many major computer manufacturers were using Murphi as a design tool.

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

With some of my NSF funding, I started a few students working on this problem. We developed a program called "Murphi" for checking cache coherence protocols. We tried it out on protocols being developed in our building, and a few in local industry (in Silicon Valley).

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

To make it the memory appear faster, computers use "caches" -- small, close memories that keep a copies of frequently-used memory locations for fast access. The circuitry moves data in and out of the cache as necessary.

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

So, I was in a building with researchers designing "shared memory multi-processors". These are machines with multiple CPUs that interact by sharing memory. Back then, each CPU was a separate chip, and these were big machines. Nowadays, all multi-core CPUs are shared memory multiprocessors.

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

At the time, it was enough to fund a couple of PhD students, some travel, maybe summer salary (professors often get paid for 9 out of 12 months. At major research institutions, they don't take the summer off. They spend it focusing on research, and get paid from research grants.)

Since the National Science Foundation is under threat, I thought I'd share one example of how NSF funding has benefited the U.S. (long) It's not the most important example. Maybe it's an average innovation out of thousands funded by the NSF.

But sometimes a story about one thing helps people understand better than a bunch of statistics. I finished my PhD in computer science and was lucky enough to get an Assistant Professor position at Stanford, which has a *really* good CS department.

US research drives healthcare innovation across the whole world. It's not just essential for the future of the U.S. Looking at it as competition with other countries is missing the main point. We're headed towards a virtual halt in progress world-wide.

Overall, coming a day after the government's plans to radically shrink the National Science Foundation, it's difficult to read this as anything other than an attempt to crush scientific research in the US. arstechnica.com/science/2025...

So, in ten years, Americans will be dying of disease that should have been prevented or treatable, and the economy will flounder because of inventions that didn't happen. But a lot of them won't know it.

Overall, coming a day after the government's plans to radically shrink the National Science Foundation, it's difficult to read this as anything other than an attempt to crush scientific research in the US. arstechnica.com/science/2025...