A potential problem in the Recorder of the new version

[burnningcat]

Dear Sir,

I use the new version of OpenSees 2.2.0, and it has a very good improvement. My model converge fantastically in this new version, although it can not converge in the version of 2.1.0!

The speed increase which is also very inspiring. But I find a potential problem of the new version: In the vesion of 2.1.0, the program will record the data in almost every step, it takes more time to run the analysis; but it means if your analysis can not converge or it has been interrupted when it is running, you still can get some results before it stops.

But now I find that the new version of OpenSees 2.2.0 will only record the output to the file when the analysis finish. It helps to increase the speed, but when something happens during the analysis, you will not know what happens during the analysis and in which state it stops.

So I sincerely ask if you can add an option to the Recorder, so we can decide for how many steps to write the results to the output file?(For example, write to the output file in 20 steps). I really like the OpenSees and I wish that it will become better and better, thank you very much!

[fmk]

the file does get written at every step .. the operating system is just not flushing the data .. if the program quits normally, i.e. without a segemnatation fault, the data will be in the file.