Dear collegues, Could you give view on StarCD and CFX 5? What's "better" and "more strongly"? Our university want to buy licence of one of those but we want to do the best choice. Our specialists and I too didn't use those CFD codes. We deal with heat exchanger tasks (especially complicated external streamline, free surface and porous solids with thermal flux). I'm impatiencing yours answers. Thanks.

Hi all,

On the posted question from Suteh a lot of answers are posted in the CFX-forum. According to these answers StarCD really sucks compared to CFX-5. Could comment on these answers, please.

Regards, Astrid

If you bothered to read closely, you would find that the guy who claims CFX is so superior in every way actually works for CFX. Do you think it is possible that maybe he is not the most objective person in the world? If you believe everything he says, then can I interest you in purchasing a bridge in Brooklyn that I happen to own.

Ahh, but there were user's comments as well, weren't there?

Yes there were some user comments and they were mostly of the sort that STAR does some things better and CFX does some other things better. Which one is better suited to you really depends on your application.

(1). In the flea market, I asked whether the 60dollar Rolex watch is more accurate than a Chinese made 5dollar quartz watch, the seller said it is a real automatic watch, and it can keep running for about four hours continuously. (2). I didn't buy the Rolex or the quartz one, because I already have two very nice solar powered Citizen watch, accurate yet no battery change required.

Of course I noticed that Robin Steed is working at CFX. For that reason I was looking for some answers from the other side to let it all be more balanced...........(or how show I say that in proper English).........

Astrid

If you had bothered to read that "guy's" post you would have read that he was working for CFX and not "completly" objective.

"I am a CFX employee, so you can take my comments with the necessary grain of salt ."

and the reason why the post is interresting is

"Before coming to CFX a year ago however, I was a Star-CD user"

Mayby there are some former CFX user who has shifted lately to Star-CD that can give more comments, even employees at Star-CD is highly welcome.

Anyway I do think that the different codes has different advantages, and one has to select the code which suits the specific needs best. In my case I do believe CFX is most suited.

Look, Astrid read some messages from someone who, by his own admission, is not objective and then declared that "STAR-CD must suck". All I was trying to do was reinforce the fact to Astrid that this comparison was written by someone trying to sell an alternate code. Of course he made CFX look better. If he said great things about STAR, he wouldn't last long as a CFX employee. The fact that he was a STAR user does not give him any more credibility. The CFX employee clearly identified himself as such and I have no problem with that.

(1). I think, I had mentioned this story of mine here before. There was this world leading company using CFX-TASCflow. They used this code to design a product, and claimed that the new product is better and engineers in the company like to use the code. (2). I tried to simulate the results, but the results showed no improvement at all. One day, I met an engineer from the headquarter and told him my story. He said that the person who did the analysis was no longer working in that area. (3). It is all right to say that you like code-A or code-B. But if you make the wrong claim world wide, it can be hard to take.

When I said that Star CD "sucks", this was meant to attract attention, not to treat Star CD and their users with comtempt. But so far hardly any objection seen. This could at least mean three things 1) Star CD users are very pleased and feel superior within their application, 2) they live on an island, 3) they don't care at all and don't want to share their experience.

Astrid

1) I have no other choice than to use STAR-CD. 2) But my opinion is that for single fluid flow with no combustion involved, with or without porous media you should decide which GUI you prefer. 3) When it comes to two phase flow or turbomachinery I would prefer CFX-TascFlow because AEA technology, the Vendor of CFX has much more experience in these fields than the others 4) For simulations with meshes changing in time STAR-CD seems to be better 5) For free surface flows you should work on your own code. STAR and CFX in general have the capibilty to solve such flows but if you are interested in phenomena which occur directly at the free surface (heat exchange, mass transfer, turbulence influences) all codes will fail because the free surface always smears over more then one cell. So the density gradient and all other property changes will not become sharp there

hello astrid,

but maybe it simply means that no one of us ever worked with cfx, so that we can't tell you which code is better (what ever this means) - and we (the users) don't want to sell this code and so there is no reason to tell you fairy-tales that (pro)star is the best code ever written or something like that.

i think that the star users share there experience like you can see in this forum - so if you have a more specific question you surely will find help here.

regards

peter

p.s.: to answer at least one part of your question:

i personaly think that (pro)star is not really easy to learn and to work with. so for 'part time' users i would not recommend it - even i don't know if there are other pre/post processors and codes which are essentially easier to use - star is the only 3d code i worked with, but i used a lot of meshing tools like icem and i see my colleagues working with several fe-pre/post processors which are compareable with respect to functionality and complexity. in my opinion they are primarily 'different' and not 'better' or 'worse'.

but for 'full time' users it is a very powerful tool and all the things seeming so old fashioned like the command line or the fortran user coding are very powerful.

where robin steed from cfx is right is that the post processor is not really up to date and makeing good movies and visualisations can be a tedious work. but the question is how important this is for you. i think that the result quality is more important than the movie quality. with each code you will get a lot of colourised pictures, and with each code you'll have to know a lot about physics and numerics to get good results and to interpret them in the right way. so in the end the question which code is easier to use and which one converges faster and is more 'stable' (by the way: don't mix up stability with inaccuracy - you can get (pseudo) 'stability' and robustness by smearing out all the problems) is not the most important point. the really interesting question is, with which code YOU can get the better results compared with meassurements or what ever your benchmark is. (i know this sounds a little bit like john, but sometimes he's simply right).

Thank you for your answer. I have only used CFX-4 (hex mesh, fortran, command lines, very powerfull, sounds like STAR CD) and CFX-5 (tets, coupled solver, no fortran yet, black box approach so don't ask what's inside, easy to use). Post processing remains miserable, probably like Star CD. I would recommend Fieldview for post processing.

Astrid

I am suprised that CFX let their employees write such unjustifiable and frankly wrong nonsense in a public forum. In my opinion it reflects badly on both the company and its product.

I've been in the industry for almost 10 years, and in that time I've used all of the big three codes at one time or another.

I've just been through a thorough code selection process in which we considered CFX, Fluent and STAR-CD.

Although I was fairly impressed with CFX-5, I am certainly not prepared to invest time and money training my engineers to use a product with a very uncertain future.

It is no secret that CFX is up for sale. It's no secret either that no-one wants to buy it. AEA want to sell CFX because it loses money, no-one wants to buy it because frankly it is a rather unattractive product compared to its rivals.

Most industries can support a market leader and a major competitor, almost always the third place company fades away to nothing.

In the last 5 years CFX has been moving backwards at a rate of knots. From a clear second place in the big three, it has been overtaken and thoroughly eclipsed by STAR-CD. So while STAR and FLUENT battle out the lead CFD is left as an also ran; a Cyrus to AMD and Intel, a Virgin Cola to Pepsi and Coke and an OS/2 to Linux and Windows.

The market doesn't lie and provides a very efficeint form of natural selection. You only have to look and see how Flomerics are forced to lay off 20% of their workforce after years and years of excellent performance.

So in my opinion you should limit your choice of code to the market leader and its close challenger. Unless you have a very good reason otherwise, I suggest you choose a product with a clear future.

Once again the punters don't lie.

Regards

William

As someone who could care less (at this point) about who's software is 'best', I do find it amusing that those who have the worst to say about CFX hide behind first names with no further contact details. In my experience, this is akin to stabbing someone in the back. As has been said before, at least those CFX employees that gave their opinions on the oppositions software did so very openly!

Well said! Seems to me like Will is working for Adapco or is narrow-minded enough to realise Start-Cd does the job okay for his particular application and this therefore makes it the BEST code in the world! In my applications of multiphase flow with chemistry there are certainly better codes than the one Will put forward. This does not make them the BEST codes ever though...

Folks, As an adapco employee There aint no william here on LI except one of the owners and he is an Ansys user! Runs our structural constulting group and wouldn't waste his time on a CFD forum.

Could be an alias, or someone from CD in London or CD-adapco-Japan. But in general we are a brutally honest little group. We know our code aint perfect cause we get bitten in much more tender places much more often than most of our users.

Some of us have been being bitten for over 10 years!

Anyway CFD is a tool. Our tool does some things well, and some things best, and somethings not at all. We would be the first to grant our competitors the same. I have an academic friend with access to all three codes and he says when a new version comes out he can see who copied what from whom.

Suteh - I would recommend you contact the sales orginization for the codes you are interested in. Provide them with a complete description of the problem ( and please not something you've simplified) and see what they think their code can do.

Quite frankly you will have no harder a time filtering out the sales staff BS than you do the BS in these lists. While a sales guy may exagerate, he wont go to far cause he wants that renewal in a year. And no John C. Chien will call.

As for why adapco/CD isn't all over this, its because not everyone has web access at work, and those of us that do have real work to do. In fact when my boss reads this response I am sure he will give me a hard time. The compiler was slow, honest Steve!

No more time to kill - Andy Robertson