Powerful Beyond Imagination

No, the title of this post isn't intended to refer to the new quad-core processors Intel announced yesterday (as I write this), though I suppose it could. Rather, it's the slogan of this year's SC06 (Supercomputing 2006) conference, whose US edition this year is being held in sunny Tampa, Florida. And yours truly is there.

This is the third Supercomputing conference I have attended, and I am astonished and surprised each year at the wide variety of exhibitors showing off hardware, software, services and projects, all in the "high performance computing" space. In what seems to be an amazing instance of Kismet, the booths of Intel and AMD, usually far apart on the show floor, are directly across from one another this year. But I digress...

What struck me, as I wandered the show floor, is that Fortran, a fifty-year-old language considered long-dead by many, has a customer base vibrant enough to support (at least) seven commercial vendors offering compilers on the same platforms, plus two (why two?) competing open source compiler projects! What other widely-adopted programming language can say the same? Some dinosaur, eh?

I'm currently reading Andy Grove: The Life and Times of an American, by Richard S. Tedlow. Grove, as many of you know, is one of the three founders of Intel and was CEO during Intel's most explosive growth in the 1990s. I didn't know too much about Grove, a Hungarian Jew who made his way to the US in the 1950s, but Tedlow tells the story well, other than the worshipful attitude of Tedlow towards Grove getting on my nerves at times.

So imagine my delight when I came across the following on page 86, as Tedlow is relating Grove's first few weeks working for Gordon Moore at Fairchild Semiconductor in 1963, sometimes quoting Grove:
Somewhere, during the course of his education, Grove had learned computer programming. This knowledge served him in good stead. He analyzed data on a batch-processing computer service that allowed him to create the "representation of that closed-form solution." [A problem Grove had been assigned to solve.] "[V]ery few people," said Grove, "knew how to program in Fortran in 1963 at a Silicon Valley commercial company."

See? Intel and Fortran go waaay back!

Also meeting this week is the Fortran standards committee. (They meet in sunny Las Vegas, but way off "the Strip".) I'll have to ask Stan Whitlock, our representative, what new things have been added to the next standard currently known as Fortran 2008. There's already a lot on the plate, and this is before a single full Fortran 2003 implementation exists. (F2003 is massive enough.) I'll talk about F2003 and F2008 in a later post.
For more complete information about compiler optimizations, see our Optimization Notice.