CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   CCM+ vs CFX vs Fluent (https://www.cfd-online.com/Forums/main/13131-ccm-vs-cfx-vs-fluent.html)

guillaume March 13, 2007 23:37

CCM+ vs CFX vs Fluent
 
Hi I am a current user of CCM+, I am wondering what CFX and Fluent offers in terms of solver speed, meshing algorythm and user friendlyness. Since they are pretty much the same price, which one is the best? I use CFD in the building industry, so I am not too worried about the turbulence models used for mach > 10...

Ben March 14, 2007 14:09

Re: CCM+ vs CFX vs Fluent
 
I have only used STAR-CCM+ so I can't talk about the user friendliness of the other two but I can't imagine they are much easier to use or much slicker (I am a big CCM+ fan). On the modelling options, for building services you will find that all the solvers have similar and equally relevant models (both physical and turbulence).

CFX differs from fluent and CCM+ in that it is a node based solver and as such has higher memory usage and in the case of transient calcs, far far slower. Also to consider is that both CFX and Fluent are going to have fairly finite life spans as Ansys is going to kill them both off and develop an entirely new solver, this of course will take years but non the less you may find in a few years time your process is thrown into disarray as you will be forced to change to a new package.

opaque March 14, 2007 17:38

Re: CCM+ vs CFX vs Fluent
 
Dear Ben,

To be able to talk about the future of two products, you must have some insider information. Otherwise, I wonder the value and intention of your comments and I consider them speculative.

Your insider contact can perhaps give you a demo license for both CFX and Fluent, and after "really" evaluating both codes your comments will regain some value.

Please be fair, and let those (real life users) who have meaningful and helpful information speak out.

Opaque


Ben March 15, 2007 03:49

Re: CCM+ vs CFX vs Fluent
 
The fluent/cfx information is from the mouth of a Fluent engineer (who will remain nameless) and so I personally wouldn't consider it speculative.

On the demo license issue I am not sure what this is likely to achieve as I am an experienced user of CCM+ and can perform everything required of me in it within the required timescales. Part of the job I am required to do is transient analysis which I know for a fact CFX is poor on (I know this from CFX users and an ex CFX engineer).

Phil March 15, 2007 13:19

Re: CCM+ vs CFX vs Fluent
 
Rubbish on all counts. CFX may have somewhat higher memory requirements, but that is because it has a coupled solver, not because it is node-based. And the coupled solver is why it has a strong reputation. But get your own experience... Ben and I are obviously both biased.

CFX does have a reputation for being slower on transient problems, but that is probably ill-deserved. I've seen a report that show CFX-10 and CFX-4 (an older fast code) to be roughly the same speed on an LES example when following best practices for each code.

CFX and Fluent will both be around for a long time - this is how Ansys makes their money, so rest assured they won't do anything to disrupt it.

Ben March 15, 2007 14:16

Re: CCM+ vs CFX vs Fluent
 
Ok my bad I forgot that CFX was a coupled solver only too, although a node based solver will inherently lead to higher memory usage as you are solving at a greater number of points than a face or cell based solver....

I also agree that fluent and CFX are going to be around for a long time, in fact CCM+ is a good example as CD's prime income is still STAR-CD and they would be mad to can it. It is just something to bear in mind that eventually the process will change eventually.

Phil March 15, 2007 14:30

Re: CCM+ vs CFX vs Fluent
 
I'm glad we agree!

I still don't quite agree with what you say about a node-based solver memory requirements though. On a hex mesh, node-based and cell-based are roughly equivalent. On a tet mesh, node-based has about 5X fewer solution points but also about 5X more neighbours per solution point, so one might expect matrix storage to be roughly similar. It will probably depend heavily on implementation details. (I'd also expect a polyhedral cell-based method to have very similar requirements to a node-based solver like CFX, since the control volume layouts are quite similar.)


Ben March 15, 2007 14:49

Re: CCM+ vs CFX vs Fluent
 
This is true and indeed one of the reasons that polyhedral meshes converge quickly (the neighbourhood connectivity). I guess it's all much of a muchness. In many ways the big three aren't going to vary much in capability and speed (they can't for commercial reasons) and it can be purely a matter of GUI preference, I guess in this regard CCM+ and CFX probably have the edge over fluent as both are far newer than fluent.


Josh March 15, 2007 15:24

Re: CCM+ vs CFX vs Fluent
 
for building I would rather use www.windsim.com and their distributors. It will beat ccm+, fluent and cfx in term of runtime and price.



All times are GMT -4. The time now is 02:17.