Depends on what you want to do. If you just want to use code to do research using available physical models, commercial alternatives are probably the most efficient. However, if you need to use custom models for physics, turbulence etc, then I would recommend OpenFOAM as an open source alternative that lets you extend any parts of the code. Also a budget alternative since it has no license fees. Quite steep learning curve though so if you don't need the added flexibility it may not be the best choice.

ANSYS CFX.

-Faizan

I don't believe there is the "best simulation software". The selection depends on the simulation task at hand and on the budget you want to spend. Without any additional information, ANSYS CFX seems to be a safe suggestion, but it's not cheap and it requires significant skill and experience.

you can use AFEPack with help of Linux this is very useful in CFD

Depends on what you want to do. If you just want to use code to do research using available physical models, commercial alternatives are probably the most efficient. However, if you need to use custom models for physics, turbulence etc, then I would recommend OpenFOAM as an open source alternative that lets you extend any parts of the code. Also a budget alternative since it has no license fees. Quite steep learning curve though so if you don't need the added flexibility it may not be the best choice.

For me in research it would be always OpenFOAM only due to the fact that you really can manipulate everything and you know what you really solve. Its an open source toolbox and powerful if you know how to handle it. In our department more and more people switch to openfoam.

That depends and envolves a lot of questions. Will it be direct numerical simulation (DNS)? LES? RANS? Are you studying isotropic turbulence? Are you seeking spectral accuracy? Is your geometry complicated? Does it envolve multi-physics? There are tons of alternatives for each case. In a more general sense, I would recommend OpenFOAM as stated by others above. It covers a wide range of flow configurations, offers multiphysics solutions and allows the user to program and modify it, being open-source and free. Keep in mind that, depending on the case you are studying, this may not be the best alternative.

It is so intereting to see such a question fluid dynamic research. What do you mean by this? are you interested in developing a code for fluid structure interaction? or it is to understand the flow properties around a given shape or hull? My answer will be a similar one to that of Leonardo and Emmanuel.

In my opinion, FORTRAN is the best program because it can handle complicated phenomena in fluid mechanics while other do not.

Dear colleagues,

    Thank you all for your valuable contributions.

Best Regards

I would recommend COMSOL multiphysics with the appropriate physical module.

If you want to do long term investigations buying the software is worth. 

Greetings

If you're really feeling ambitious, make your own. Not easy, but you'll get tones of practical knowledge, and you'll know what your simulation's doing. Failing that, openFOAM is pretty good

I think that COMSOL Software might be also interesting

In my opinion I would like to suggest a software by ESS called SENSE revolutionizes the CFD-simulation market It is really easy to use and accurate particle based solving running on cost-efficient GPU,