Message boards : Number crunching : Language used.
Author | Message |
---|---|
adrianxw Send message Joined: 18 Sep 05 Posts: 653 Credit: 11,840,739 RAC: 28 |
I thought I'd asked this before but can't find the thread? <fx>shrugs</fx> What language(s) are the applications written in, and if relevent, what version(s). Are there any plans to change the current choice(s)? Wave upon wave of demented avengers march cheerfully out of obscurity into the dream. |
dcdc Send message Joined: 3 Nov 05 Posts: 1832 Credit: 119,668,410 RAC: 10,687 |
I believe rosetta beta was fortran and minirosetta was the product of converting this into C++ (or maybe C?) |
adrianxw Send message Joined: 18 Sep 05 Posts: 653 Credit: 11,840,739 RAC: 28 |
Yah, I'm kind of looking for a more precise answer, Fortran-IV, Fortran-66, Fortran-77, '90, '95 etc. Wave upon wave of demented avengers march cheerfully out of obscurity into the dream. |
Murasaki Send message Joined: 20 Apr 06 Posts: 303 Credit: 511,418 RAC: 0 |
|
Mod.Sense Volunteer moderator Send message Joined: 22 Aug 06 Posts: 4018 Credit: 0 RAC: 0 |
I believe the Fortran was longer ago then any of the versions in use today. And prior posts have indicated code is in C++. ...why do you ask? Rosetta Moderator: Mod.Sense |
casio7131 Send message Joined: 10 Oct 05 Posts: 35 Credit: 149,748 RAC: 0 |
as someone who currently programs in fortran, i have some questions: 1) why the change of languages? because fortran was seen as "too old"? or because more people knew c++? 2) how long did it take to convert from fortran to c++? i suspect it would've been "longer than expected". 3) was the c++ version faster than the fortran version? i suspect the fortran version should've been much quicker (i'd say ~50%, as a rough guess). in the off chance that c++ was quicker, i'd say that would be because the c++ compiler was much newer than the fortran complier. i understand if you can't answer all (or any) of these questions, but i'm most curious about questions 3 and 1. |
adrianxw Send message Joined: 18 Sep 05 Posts: 653 Credit: 11,840,739 RAC: 28 |
>>> why do you ask? A complicated question. I programmed in Fortran-77 for 20 years, late '70's to late '90's, the last 10 years or so, I've been almost exclusively C++. My employers took a decision to migrate to C#. I really dislike it, well, the language itself is not unlike C++ but with many really dumb features designed to "make it easy" for beginners, thus encouraging more prople into the field. What it actually does is clouds the computer from the programmer, after all, "programmers don't need to know about the machines they work on do they" ethic seems to have been applied, excessively. Time to move on. I wondered about my Fortran history. I know 77 is used a lot, but wondered if it was "worth" learning the updates, 90, 95 you know. I asked here, and at a couple of other science projects, because I fancy doing more scientific stuff rather than the deeply embedded engineering stuff I've been doing for best part of 30 years. That kind of starts the reason for my question anyway, I could go on, {and on, and on...} casio's question's, 3 in particular, concerning performance is interesting to me as well. Wave upon wave of demented avengers march cheerfully out of obscurity into the dream. |
Mod.Sense Volunteer moderator Send message Joined: 22 Aug 06 Posts: 4018 Credit: 0 RAC: 0 |
casio, understand that I have no first-hand knowledge of the details. But I believe I'd be correct to say that the primary reason for transitioning from Fortran to C++ was that more people understand C++. The code has contributors from labs around the world (Rosetta Commons mentions about 100 contributors!), and so I believe C++ was a language that many are already familiar with. And with the constant enhancements and new protocols being added, I believe the OO aspects probably were beneficial as well. No idea on any performance differences. Performance doesn't matter if you don't have a program to run, and by using a more familiar language, the field is more wide open to incorporate methods and protocols that others are working on that you may not have otherwise. Rosetta Moderator: Mod.Sense |
casio7131 Send message Joined: 10 Oct 05 Posts: 35 Credit: 149,748 RAC: 0 |
thanks for your response mod.sense. i guess the change to c++ makes sense (pun not intended) when you consider the number of contributors/users of the rosetta code. adrianwx: if you're going to do some numerical stuff, i'd say that learning f95+ is quite useful - i'd learn f95 and maybe use some f2003 features (most compilers support much of the f2003 standard now). the array operations, modules, allocatables, etc. make programming much easier and safer than f77. also, fixed-form source code is still valid in f90+ (but free-form is usually preferred for new code) so standard-conforming f77 code should be able to be used in f90+ with virtually no modification. |
Mike Send message Joined: 30 Apr 09 Posts: 44 Credit: 65,019 RAC: 0 |
the graphics is openGL certainly |
Message boards :
Number crunching :
Language used.
©2024 University of Washington
https://www.bakerlab.org