Last month, there was wall-to-wall media coverage of how the US was back at the top of the supercomputing leagues for the first time in almost three years, reclaiming the title from Japan, thanks to a new system developed by IBM.

The computer, Sequoia, uses IBM’s BlueGene/Q servers and at the US Department of Energy’s Lawrence Livermore National Laboratory it achieved 16.32 petaflops – trillions of floating-point calculations per second – using more than 1.5m processor cores.

For many of us, it was a time to remember the time, thirty years ago, that the US pioneered the use of parallel supercomputers, and in particular to remember the pioneers themselves such as Ron Gruner.

In 1982, the computer engineer Ron Gruner, together with Rich McAndrew and Craig Mundie, founded Alliant Computer Systems to build parallel supercomputers, the aim was to develop a machine which could achieve better performance than the fastest single CPU-machines and in a way that was transparent to developers.

After three years of work, Alliant had cracked how to do it; for several years the company was one of the leading manufacturers of supercomputers in the world. Then, it lost its way, was late on shipping the next generation models and, like many others of its kind, Alliant went into administration in the late 1980s.

Ron Gruner then began to think of how to build a company with recurring revenue and first looked at telephone technology, ancient history now, but back in early 1990s it was really hot. By 1992, he founded Shareholder.com with the aim of using technology to automate investor relations. Shareholder grew and was acquired in 2006 by NASDAQ.

RM:

Ron how did you get involved with computers? Was it anything to do with the Homebrew clubs which were very popular in the US?

RG:

Well, it all started with electric trains. I had a large Lionel layout when I was a kid and built a system of relays to allow three trains to run automatically without collisions. That got me interested in logic design which naturally led to computers.

RM:

Obviously there were fewer computer books available then than there are now, but are there books that you found particularly useful or books that you think programmers should read? Have you read Knuth?

RG:

There were actually many computer-related books in the late 1960s. Donald Knuth had just published The Art of Computer Programming. I was a hardware guy, so didn’t really study his book. My bible was Arithmetic Operations in Digital Computers by R.K. Richards. It was written in the 50s in the time of the vacuum-tube computers, but provided a clear description of binary arithmetic. In the early 70s Gordon Bell and Allen Newell published Computer Structures, a book I found very useful.

RM:

Was there anything that drew you into computers other than the thought ‘I seem to be good at this”?

RG:

Yes, at the time I was a Ham Radio Operator transmitting at typically 5 megahertz. I remember being just amazed that a computer operated at the same frequency, but in one second performed five million discrete operations, an add or branch decision for example. People take it for granted now, but anything that can divide time into a millionth, or nowadays a billionth, of a second and then do complex operations in that time was just astonishing to me. And, of course, there was HAL from the film ‘2001, a Space Odyssey’ who we early geeks worshiped even though he did develop a mean streak.

RM:

It sounds as though you viewed computer programming as a means to get neat things done, did you learn to program in order to make things happen? What languages were you programming in?

RG:

I was a hardware guy until the early 1990s. By then the computer wars were over and the desktop had clearly won, so there just wasn’t the same need for computer and logic designers that existed in the days of mainframes and minicomputers. I started programming with the Internet in HTML, ColdFusion and SQL. I loved it from the beginning because, unlike hardware design where you had to wait months to have your designs actually built, programming was immediate. You had an idea, you designed and programmed it, and voila it was done and working. The other nice thing about programming is that it’s pure logic. With hardware, you have to deal with physical issues such as signal coupling, heat dissipation, propagations times and such. All this can be fun too, but they’re still secondary to the algorithm you’re implementing.

RM:

You spent some time designing computers for Data General who seemed to have a remarkable take on how to incentivize its staff. I understand they had two teams which competed against each other to get products out on time. Did you find that useful in any way?

RG:

Since I was on the losing side of the Eagle/FHP wars documented in “The Soul of a New Machine” it’s not too surprising that I found the approach of dividing the company into warring camps to be destructive. In the three companies I’ve started since Data General I’ve tried to focus our people on innovation, client satisfaction and the simple pride of a job well done. Sounds a little dopey, but these are terrific motivators for people.

RM:

It’s certainly an interesting way of building profit! You’ve hired plenty of people now and in the past: How you recognize the talented ones? Microsoft is famous for asking puzzle-type questions. What do you think of that as a way of gauging someone’s potential?

RG:

In my experience finding really smart people is the easy part. There are lots of them around. I always looked for people who had a strong work ethic, worked well with other people and were curious. Curiosity generates innovation. Working hard in teams executes that innovation. There are exceptions. Over the years I’ve hired my share of prima donnas, but they’re the exception.

RM:

You left DG in the early 1980s and together with Craig Mundie and Rich McAndrew founded Alliant Computer Systems to build parallel supercomputers. What inspired to go for this market? What were the big turning points for the company?

RG:

Supercomputing was a hot technology in the eighties, being thousands of time faster than minicomputers and the early personal computers. They promised to open up all kinds of new applications such as simulating complex chemical interactions, image processing and aircraft design. Yet, being built from discrete components, their speed was limited by the speed of light, since signals typically had to travel several feet, not micro-inches as in today’s microprocessors. So the idea was to decompose a program and run the various parts in parallel. We discovered that Dr. David Kuch at the University of Illinois had done promising research on program decomposition. We met with him and signed an agreement to commercialize his research. That was the basis for Alliant.

RM:

In the early days of Alliant, was your life an endless series of late nights, sleeping late and overall working long hours? How much of that is a necessary part of being inside a start-up?

RG:

We knew that other start-ups such as Convex Computer were also moving into this market, so time to market was very critical. We told people when they were hired that every other Saturday was a standard work day. That was just a start. Over our two and a half year development period, our facility was only empty three days. We were all young, many without families, and worked incredibly hard. Today, with home computers I think people work even harder.

RM:

How often were there design reviews? Was it at the beginning of a project, or throughout?

RG:

We had design reviews continually. Although the Alliant system was very complex, by the mid-1980s we had good hardware and software emulators. So when the actually computer hardware was finally assembled, the entire system worked perfectly because the compilers, operating system and hardware (e.g. gate arrays) had already been tested extensively.

RM:

Alliant had huge success for nearly a decade but then it filed for bankruptcy protection in 1992 a year after you left the company. Was its undoing due to it not being able to move quickly enough to embrace change?

RG:

As the CEO I made several serious mistakes which helped sink the company. We got into a disastrous relationship with our Japanese distributor, we missed a major product cycle, we didn’t cut expenses quickly and deeply enough. The late eighties were a tough time for the traditional computer companies. The performance, flexibility and low-cost of workstations and later personal computers just blew this market apart. Very few companies survived.

RM:

Do you consider your work at Alliant one of your important technical achievements?

RG:

In terms of helping put together a fantastic team of people along with my co-founders Mundie and McAndrew, yes absolutely. Looking back 30 years, it still amazes me what we accomplished. At a personal level, my greatest technical contributions were in the seventies when I designed the Nova 1200, Eclipse and major parts of the FHP system all while I was at Data General. Today, although it’s not as technically challenging, I’m really proud of the work we’re doing at Sky Analytics to help legal clients benchmark and manage their outside legal spending. This is a multi-billion dollar problem and we’re providing the tools to help solve it.

RM:

After Alliant, you set up Shareholder.com which generated income through investor relations and this was before the Internet really took off with corporate America. When the internet did come along how hard was it do the missionary work with your clients to make them feel safe about the technology you were using?

RG:

I had my Internet epiphany in 1994 when I traced a FedEx shipment over the web. A few months later I went to the president of the National Investor Relations Institute (NIRI) and gave him a web demo. He was astonished. I told him this was going to change everything in shareholder communications. We then offered to provide free web design and hosting for NIRI as well as three other industry trade associations. These associations then became advocates helping evangelize the benefits of Internet-based shareholder communications.

RM:

If I may ask you a couple of random questions Ron. Would you say that the computer industry has been invariably wrong about what computing is and where it should go and how to improve? For instance if you look at the role of ARPA and contrast with what industry thought it should be doing, it was way off the mark. Is this situation any worse today?

RG:

Richard, I don’t completely agree with your comment regarding ARPA. They certainly played a key role in Multics which seeded much of today’s computer security technology and the Arpanet which was the foundation for today’s Internet, for example. Regarding today, I am worried that we’ve come full circle. Thirty years ago every computer manufacturer had its own proprietary operating system and peripherals. Later when email developed, all the systems were incompatible. It was a mess. Then Unix became a unifying operating system and the Internet provided a simple, universal system for all of us to communicate and build applications. Now with Apple’s iOS, Google’s Android and Microsoft’s Windows 8 we’re moving back to closed, proprietary systems. It’s a discouraging trend for the long term health of computing in my opinion.

RM:

In what way?

RG:

All these proprietary systems make it much more difficult to develop universal products. For example, I’ve developed a web-based flight calculator used to calculate flight parameters for small jets. That worked great, but was tied to the Internet. We then developed a stand-alone version for the iPad. That turned out to be a huge job and took six times as many lines of code even with the help of a highly experienced iOS programmer. Now there’s demand for an Android based version. These are all largely independent development projects. Of course, Apple brought a huge amount of innovation to the user interface, so the extra effort was worth it. My guess is that over the long term we’ll see a convergence making it simpler to write applications for multiple platforms.

RM:

Do you think many people working in technology are unaware of its history and have little curiosity about where languages for example came from?

RG:

That’s probably true, but let’s remembers that most people today grew up with computers and so pretty much take them for granted, even though they are truly miraculous machines, the most complex mankind has developed. I didn’t think much about electric motors when I was a kid, but I’m sure they were fascinating when they were a new invention.

RM:

You were in at the start of parallel hardware: Once we do really know how to exploit parallelism, will there be any need for high-end processors, because we will be able to get the same performance with much greater energy-efficiency by using larger numbers of simpler processors?

RG:

I think this question is a little off the mark. Today all high end processors are parallel machines constructed from thousands of small, simple microprocessors. Nearly all compute intensive applications can be decomposed into parallel operations which can run simultaneously although ultimately limited by Amdahl’s Law.

RM:

When you look back at your career on all the things you have done is there one time or a period that stands out among all the others?

RG:

Yes, now! The tools to which we all have access — incredibly fast, easy-to-use, cheap computers; mobile devices and their apps; great software development tools; the Internet — say to me we’re in the golden age of computing. All these tools make development cycles much shorter which is a lot more fun that waiting months or even years to complete a project.

RM:

Would you consider yourself a scientist, an engineer, an artist, a craftsman or something else?

RG:

I’ve always thought of myself as an engineer. Engineers build things. That’s what I’ve enjoyed most about my professional life. Building things.