Hi,

I try to plot residuals from Prostar with the button "Plot residual history" but after some time from the start of the run, it does not work anymore !? Prostar reports an error in the Case.rsi file. So it works initially but not later on which points towards a problem that some parameter needs to be increased (?) Does anybody know a solution to this !

Why don't you just increase the parameter ProSTAR tells you to? Just use the command MEMO.

Star does not tell me to increase any parameter, that is the problem. It reports "ERROR - Invalid file Case.rsi chosen. Format does not correspond to Star residual file." This seems to be actually a operating system error, since this is the first time I am using Linux. Previously I have never obtained this kind of message.

Perhaps this is a problem of the Linux-version. I got the same problem sometimes, but sometimes it worked fine with the same case. On Unix, it never happened. Did you ask the support?

Regards,

Nicola

Hmmm, not sure about .rsi files but I could open .mdl, .pst, and even use binary .geom and .prob files taken over from IRIX, HP-UX and Windows OS in Linux.

What OS did you get your Case.rsi from?

Linux, Red Hat 8.0, installed in pc.

Hi Ossi, try convert residual data from binary to coded form, then add blank line at the end of the new file and finally convert this new file to binary form. It should help you.

Vladimir

After the run has finished, there was no problem to open the .rsi, but what if I want to check the residuals during calculation? There was no problem to open any file transported from Unix to Linux, neither to start a unix.mdl on a Linux Red Hat 8.0 PC. I have difficulties only in online monitoring residuals.

Nicola

I had the same problems and solved it the way I described, I thing at least. If you want to look at residual values only, use .run file or look at a terminal which you used to start Prostar, otherwise make a copy of your actual .rsi file and do what I said in previous message with that file copied. Unfortunately, to be able to use my procedure, you should find out something about commands used for displaying of the residuals. (commands : GRES, GSTA, GDRA ...).

Regards Vladimir

This procedure should work if the job is still running:

Graph menu --> Load graph data dialog box: Load star residuals and monitored values next box: choose desired .rsi file, go down to iterations and change "all" to "range." For the range, put the max iteration to be one less than the last iteration done. For example, if the job has completed 242 iterations, then the range should be 1 to 241. Hit load data, then graph data. The residuals should appear in the graphics window.

You can also change the Values to extract button from "both" to either "residual" or "monitored."

Hi Vladimir,

Sounds useful. But what command do I use to convert the binary .rsi files to coded?

Has anybody gotten STARWATCH to work on Linux before?

If you have a big screen, you can just watch the residuals as they are printed on the screen real-time...

It is somewhere in the menues. Sorry, that I do not tell you the commands exactly, but I do not have access to graphical interface now. If I remember, it should be at the the end of file menu or better look for menu called 'convert'.

regards

Vladimir

Right, found it. SMCO.

Hello Ossi,

I get this message (HP-UX) if I try to plot the residuals or monitoring data while the .rsi file is being updated. If I wait until the file writing is done, the problem is gone as well. Of course, it can be annoying once in a while with small models where the residuals are written every other second, but I don't really see this as a problem. If you absolutely want to check your values in the middle of the .rsi dump you can always have a look in the .run file.

Lars Ola

I've had this problem a lot when doing a coal combustion analysis. One can increase the amount of residual one can plot by increasing maxstor. I belive the problem to be that the residual file does not get closedproperly, whatever that means but if one uses the command: gstar,casename.rsi,flui,0,0,0,resi,1,n-1,1,n where n is the iteration the solver is on, and then gdraw 1,2 then the residuals will be plotted up to the penultimate iteration.

allan