Crashing before writing files...

[James Roberts]

I'm busy playing with a fairly extreme melt composition derived from bulk summation modelling. I'm looking to compare the first solids formed to the actual rock succession, under a variety of oxidation conditions. However, the program keeps crashing before writing the tables- I can read the screen output (which looks promising for the experiment) but the lack of a final table is killing. I assume I'm doing something wrong...

Anyhow, here's the input variables:
Title: Roberts 2014
Initial Composition: SiO2 41.52
Initial Composition: TiO2 2.79
Initial Composition: Al2O3 15.51
Initial Composition: Fe2O3
Initial Composition: Cr2O3
Initial Composition: FeO 19.72
Initial Composition: MgO 4.05
Initial Composition: CaO 9.14
Initial Composition: Na2O 3.8
Initial Composition: K2O 0.38
Initial Composition: P2O5 0.64
Initial Composition: H2O
Initial Temperature: 1400.00
Final Temperature: 1000.00
Increment Temperature: 5.00
Initial Pressure: 1.00
Final Pressure: 1.00
Increment Pressure: 0.00
dp/dt: 0.00
log fo2 Path: None
Mode: Fractionate Solids

I'm using the frac_xtal_env.txt file, and a VM Linux installation. When I set the fO2 buffer to FMQ, the program runs for a bit but terminates early. Increasing the pressure to 1000 bars resulted in a complete crash, as did playing with the oxygen buffer. Any thoughts?

James

[Paula]

Hi James,

I'm sorry, I just realised that I never replied to your post. It was made when we had barely rung in the new year here!

So, I haven't been able to reproduce the problem... which is to say, the program does crash as you describe but it always writes out the output files up to the step before the crash (which is way after anything meaningful is happening, see below). So it is likely that the files are being written but just not where you are expecting them to be.

How are you starting run_alphamelts.command? If you are double-clicking the link then the output files will be written to the folder where the links are or, on Mac, to your home space. You can change this behaviour by typing '-p ' followed by a folder name when it asks for command line switches -- you can drag and drop the folder from a file browser like Finder or Windows Explorer. Note that, if the program does not crash then you need to exit (menu option 0) before the output files will be processed and moved to the folder that you chose with '-p'.

It runs till about 1000 deg C, at which point TiO2 is exhausted. The crashes you are getting are to do with round offs / negative mole fractions of TiO2 (sometimes of Fe2O3) etc. As you say, the composition you are using are quite extreme, especially when you have no fO2 buffer and have no Fe2O3 in the bulk. With those conditions there is, apparently, not a suitable solid phase to put the sodium in and at the point when the calculated stops, with or w/o segmentation fault, there is about 12 wt% Na2O in the liquid. Without knowing what you trying to achieve with the calculations, I'm not sure what to suggest...

Hope that helps and sorry again for the delay.

Paula