Hi, I've an ACCESS VIOLATION error while openning a .dat file and I don´t have any idea why it happens.My computer is a windows XP machine.

This .dat file comes from a windows 2000 PC where it works well.

I´ve copy the .cas and .dat files by FTP. (I've not select the "write binary file" when I saved the files, in order to obtain ascii files).

I´ve looked for a similar error in the Fluent forum but I haven´t found anyone....

PS: I have the same problem when I try to init the case....

Can anyone help me please???

Thanks

Does the case read fine?

You said that the ftp transfer was done in text mode. As I know, implicitly, fluent saves the .cas and .dat files in binary format (usually it takes less space) so probably this is the source of your problem. I suggest allways to use the binary mode when transfering through ftp. You can also try to open with notepad the files on both machines (XP and win2000) and see if they are realy in text format.

Dragos

I think Fluent ASCII files should also be transferred as binary. Check.

I've transferred the files as binary.

I don't have any problem reading the case file, only a warning:"process threading not available in this solver."

I don´t know what to do......

This may sound naive, but are you using the same fluent version on both computers?

Yes. (It is not a naive question..)

Were you by any chance running Fluent Parallel in the Windows 2000 machine? I suspect the Win 2K machine could be a dual-processor one.

No, i don't running in Fluent Parallel

Can you send me the case and dat files?

I've found it!!!! It was a dummy error: I'had wrote "3d" instead of "3ddp" in the file: user_nt.udf

Thanks pU1| and Dragos

This brings me back to a pointer regarding choosing the solver type (2d, 2ddp, 3d, 3ddp) for those who have a case file already. Start fluent from the command line without specifying any solver type. I am sure this can be done in *nix systems:

[user@HOST dir]$ fluent &

Once this is done, directly read the case file without specifying the solver. Fluent will determine automatically which solver was used in conjunction with the case being read and load it prior to reading the entire case. This way problems such as the one outlined in this thread can be avoided.

Best Regards,

Srinath Madhavan

Correction: I do not think that DavidSFs' problem is a correct example for the pointer I just mentioned.

Now, I am curious, what is this file user_nt.udf. I always work with *nix systems and have faced no need to ever edit any makefiles or other cpp generated files. Can anybody shed more light on the nature of the user_nt.udf file?