CFD's three pillars in math, computer science and mechanics

pyroknife · January 21, 2013, 02:42

On wiki, it stated CFD's three pillars are in math, computer science and mechanics.

How true is that?

cfdnewbie · January 21, 2013, 02:44

It is true ....what's the reason for your question??

pyroknife · January 21, 2013, 02:48

Quote:

Originally Posted by cfdnewbie

It is true ....what's the reason for your question??

I was just curious. I'm an undergrad considering going into CFD. Those areas are my stronger areas.

cfdnewbie · January 21, 2013, 02:50

you will need a solid education in fluid mechanics then, plus some good math skills. and don't be afraid to program, and learn C or Fortran. Good luck!

pyroknife · January 21, 2013, 03:25

Quote:

Originally Posted by cfdnewbie

you will need a solid education in fluid mechanics then, plus some good math skills. and don't be afraid to program, and learn C or Fortran. Good luck!

Thanks. I love fluid mechanics, it's definitely my fav area.
Isn't matlab conmonly used in CFD?

You said C 'or' fortran. Is learning both better or would just one be enough?

cfdnewbie · January 21, 2013, 04:12

Matlab is a very useful tools for trying out new stuff in 1 (or max. 2) dimension, and it is nice to do quick visualizations and such. But it is no good for real 3D.
Matlab is a smal model turbine that lets you try out different configurations and learn the basics, but if you want to propel a fighter, you need a real engine

Either Fortran or C is fine, research codes tend to be written in Fortran (exceptions exists, of course), while commercial codes are more C based. But the important thing is that you learn a real programming language suitable for HPC!

pyroknife · January 25, 2013, 18:43

Quote:

Originally Posted by cfdnewbie

Matlab is a very useful tools for trying out new stuff in 1 (or max. 2) dimension, and it is nice to do quick visualizations and such. But it is no good for real 3D.
Matlab is a smal model turbine that lets you try out different configurations and learn the basics, but if you want to propel a fighter, you need a real engine

Either Fortran or C is fine, research codes tend to be written in Fortran (exceptions exists, of course), while commercial codes are more C based. But the important thing is that you learn a real programming language suitable for HPC!

Hey, you posted in my thread:CFD's three pillars in math, computer science and mechanics

Sorry, I forgot to check back.

I am planning on doing research in CFD for grad school. I saw that you stated Fortran is mainly used for research codes and C is mainly used for commercial codes. From what I understand, research in CFD is revolves more around generating the code rather than a pre-generated (commercial?) code's application

Is knowledge of both languages (Fortran and C) necessary? Or is knowing one of the 2 sufficient for research based projects?

A professor at a lab I want to work at says they use Fortran and C, but I'm not sure if that implies both or one of the 2.

Thank you!

cfdnewbie · January 26, 2013, 06:11

One language should be sufficient, once you understand the concept of procedural programming, you should be fine. I prefer Fortran, but that's just a personal opinion.

pyroknife · January 27, 2013, 23:12

Quote:

Originally Posted by cfdnewbie

One language should be sufficient, once you understand the concept of procedural programming, you should be fine. I prefer Fortran, but that's just a personal opinion.

It wouldn't hurt knowing more. I know some university labs prefer Fortran/C++.
Would it take long to learn C++ if you already know C?

In a class I'm taking this semester, they will be teaching some language called PICBasic. I'm not sure how similar this is to BASIC. I heard Fortran should take too long to learn if you know BASIC.

January 21, 2013, 02:42	CFD's three pillars in math, computer science and mechanics	#1
pyroknife Member Join Date: Jan 2013 Posts: 39 Rep Power: 13	On wiki, it stated CFD's three pillars are in math, computer science and mechanics. How true is that?

January 21, 2013, 02:44		#2
cfdnewbie Senior Member cfdnewbie Join Date: Mar 2010 Posts: 557 Rep Power: 20	It is true ....what's the reason for your question??

January 21, 2013, 02:50		#4
cfdnewbie Senior Member cfdnewbie Join Date: Mar 2010 Posts: 557 Rep Power: 20	you will need a solid education in fluid mechanics then, plus some good math skills. and don't be afraid to program, and learn C or Fortran. Good luck!

January 21, 2013, 04:12		#6
cfdnewbie Senior Member cfdnewbie Join Date: Mar 2010 Posts: 557 Rep Power: 20	Matlab is a very useful tools for trying out new stuff in 1 (or max. 2) dimension, and it is nice to do quick visualizations and such. But it is no good for real 3D. Matlab is a smal model turbine that lets you try out different configurations and learn the basics, but if you want to propel a fighter, you need a real engine Either Fortran or C is fine, research codes tend to be written in Fortran (exceptions exists, of course), while commercial codes are more C based. But the important thing is that you learn a real programming language suitable for HPC!

January 26, 2013, 06:11		#8
cfdnewbie Senior Member cfdnewbie Join Date: Mar 2010 Posts: 557 Rep Power: 20	One language should be sufficient, once you understand the concept of procedural programming, you should be fine. I prefer Fortran, but that's just a personal opinion.