Computers Everywhere

Martin Grant, Dean of Science

Martin Grant, Dean of Science

There has been at least one unnoticed revolution in modern science: computers. When I was a young assistant professor in the 1980s, physicists suggested that we further split the discipline from our self-conscious dualism (theorists and experimentalists) and add a third category of computer modelers. With hindsight this seems quaint. Computational work has superseded any self-conscious categorization, and is now completely integrated into the work of all theorists and experimentalists, and indeed the work of all scientists.

In the 1980s, a couple of things showed me what was at the heart of scientific computing. My group had some money to buy a fancy computer, so we shipped a test code to different companies to see how fast they would run—the faster the better, of course. On one platform, recently declassified from military use, our code was reported to run blindingly fast. So we sent them another code, and waited and waited and waited for results.

Finally, the company sheepishly confessed that they couldn’t run it—and to run the previous program, they had rewritten our code in machine language. This was almost the equivalent of giving computer instructions by connecting solid-state components with a soldering iron. The compiler that turned our FORTRAN code into computer commands did not work on their computer, and I bought a different computer for my group.

Shortly after this, I was invited to a conference at a fancy new supercomputer centre, running a fancy new supercomputer. It was a very fancy place. We were all put up in a four-star hotel, wined and dined, and there was a big sushi blowout party at the end of the conference.

It was quite something. But I noticed that speaker after speaker described their work on CRAY supercomputers, and no one had used that type of supercomputer at the new centre. The centre’s director offered all of us the use of their fancy supercomputer. But every one of us who tried to run our FORTRAN codes had the same problem: the compiler did not work. Shortly thereafter the fancy supercomputer centre closed.

Finally, for me, the lesson sunk in: real scientists program on real computers, with really stable and really reliable FORTRAN compilers, which give really fast compiled code. In many ways, the people who write those compilers are the unsung heroes of our scientific revolution.

Recently, we honoured one of those heroes with an honorary degree at convocation. Frances Allen has a reputation for writing compilers that are slow, stable, and produce very, very fast— the latter are words to warm the hearts of computational scientists— compiled code. Fran’s research laid the foundation for modern optimizing compilers and automatic parallel execution.

She became an IBM Fellow in 1989—the first woman to receive the honour. In 2006, Fran became the first woman to win the prestigious A.M. Turing Award, the Association for Computing Machinery’s most prestigious technical award. To this list, McGill was proud to add the degree, Doctor of Science, honoris causa.

From computer scientists to medical researchers, from astrophysicists to undergraduate researchers, we all need the technology that Fran Allen helped create.

Comments are closed.

Blog authors are solely responsible for the content of the blogs listed in the directory. Neither the content of these blogs, nor the links to other web sites, are screened, approved, reviewed or endorsed by McGill University. The text and other material on these blogs are the opinion of the specific author and are not statements of advice, opinion, or information of McGill.