39 - HPC Cafe 2022-10-11: Modern Fortran [ID:45312]
50 von 738 angezeigt

So I'm going to talk about modern Fortran, which could be on one hand considered a contradiction

in itself, on the other hand, since it's a kind of a parodic sort of use, it's also

a possible question whether it's a future proof language or not, or something that you

can actually use also, not only today, but also in the next

decades. Given the drawing I have here, Roger has been around

for 65 years now. So it's going to have 65th birthday this year.

It was published in 1957 for the first time by John Bacchus.

Basically, it was at that time the first portable compiled

language. Similar development was for Google at such the same

time, but that was an entirely different segment of the

industry.

Different than that, Borscht's dinosaurs are still alive. So

there seem to be two main reasons. This specific dinosaur

is still in use. There are communities where Borscht is

used, focused of course on data intensive numerical calculations

in various scientific areas of endeavor, but also of course

partly in the military scientific complex, which

certainly has an impact on the continued use of the language.

Another item that is worth mentioning is that the effort of

learning the language versus the usefulness you can expect from

it. This ratio is pretty good compared to other languages.

I mean, C++ is a very, very powerful language and some

things is more powerful than Forthran, but learning the

complete language is rather a chore. For Forthran, while the

language has grown, it's not quite as bad. Typically, you

will only learn a subset anyway, but it's a much larger subset of

a complete language than in C++. Yeah, the second item is

that the dinosaur actually has adapted over time. So Forthran

is similar to C++, not the language that's been defined

once and then kept the same way forever. And in fact, the

language evolution is not driven by one, but by two

committees, which at first sight might sound a little bit

strange, more bureaucracy, does that really deliver better

results? This had actually historical reasons. When Forthran

90 was developed, there was a big, big quarrel about what to

do and what not to do. And so this led to this work working

versus implementation balance that's observed nowadays. So

the future definitions are done basically by the International

Standards Committee, while the technical implementation of the

standard is done by the US. Yeah, the other thing is, of

course, to bring the news to the masses, you also need technical

literature. So the technical literature has done its best to

keep up. And indeed, for the last decade or so, the numerous

titles sporting modern Forthran in the name have cropped up. In

naked self interest, I put out the this book especially large

because I mean, I got it the honor to, to work on the next

edition as a co author on this effort. So when the next book

appears, five. Yeah, the question of course is what

nowadays is modern about Forthran? Yeah, of course, there

are these cosmetic innovations. Those of you who know the

Teil einer Videoserie :
Teil eines Kapitels:
HPC Café

Zugänglich über

Offener Zugang

Dauer

01:07:38 Min

Aufnahmedatum

2022-10-11

Hochgeladen am

2022-10-24 21:26:23

Sprache

en-US

Talk by Dr. Reinhold Bader (LRZ Garching) at the NHR@FAU HPC Cafe, October 11, 2022

For 65 years, the Fortran programming language has been in use by scientists for writing portable and hardware-efficient code when implementing simulations and the underlying algorithms. The talk provides an overview of new developments in Fortran standardization and tries to answer the question posed in the title. We specifically focus on program design, parallelism, and interoperability.

Slides:  https://hpc.fau.de/files/2022/10/HPC-Cafe-Erlangen_ModernFortran.pdf

Tags

HPC hpc-cafe Fortran
Einbetten
Wordpress FAU Plugin
iFrame
Teilen