Perple_X 07 (Perfe_X?)

This page describes an out-of-date version, the current version is described at perplex_66.html.

Contents

What's different about Perple_X 07?

From December 2006 through February 2008 a beta version of Perple_X 07 was available as an alternative to the ostensibly stable version of Perple_X 06. As of March 2008 Perple_X 07 is the default version of Perple_X. An unfortunate consequence of the change in versions is that most of the examples and tutorials referred to at the documentation page are out of date. The out-of-date examples and tutorials are, in most cases, useful because the basic information passed between the various programs has not changed although many of the prompts have been eliminated or reordered.

The major changes introduced in the 07 version are:

• Compositional resolution and the resolution used for the mapping of phase relations is controlled by keywords (auto_refine, iteration, initial_resolution, grid_parameters, and increment) specified globally in a file named perplex_option.dat. To use the Perple_X effectively it is essential to understand these keywords.

• By default, calculations involving solution phases are done in two cycles by the program VERTEX, an exploratory stage used to estimate the range of phase compositions and an auto-refinement stage in which these estimates are refined. The two cycles are executed automatically by VERTEX if the auto_refine keyword is set to "automatic". If the auto_refine is set to "manual", then VERTEX must be executed once for each cycle.

• Perple_X uses adaptive (Connolly 2009) or non-adaptive (Connolly 2005) optimization strategies depending on the type of calculation, formerly Perple_X used the non-adaptive strategy for all calculations. In non-adaptive optimization the possible compositions of solution phases are represented by "pseudocompounds"  that are statically defined at the outset of a calculation, whereas in adaptive optimization the pseudocompounds are iteratively redefined to increase accuracy. The adaptive strategy is used for constrained optimization problems, i.e., gridded minimization typically used for isochemical phase diagram section and phase fractionation calculations. The non-adaptive strategy is used for unconstrained optimization problems, typically phase diagram (Schreinemakers) projections, multidimensional composition diagrams, and iso-potential phase diagram sections.

• Output generated from constrained minimization calculations is plotted with PSSECT and analyzed with WERAMI, both programs take the name of the problem definition file, generated with the program BUILD, as input (rather than the plot file name as in previous versions). Output generated from unconstrained calculations is plotted with PSVDRAW, as previously PSVDRAW takes the name of the plot file as input.

• The program MEEMUM does free energy minimization as a function of interactively specified environmental and compositional parameters for a system defined from a problem definition file generated with BUILD.

• Solution model file (solution_model.dat) issues:

  • Perple_X 07 cannot read certain types of microscopic formulation solution models that were included in earlier solution model files such as newest_format_solut.dat, these models must be reformulated as macroscopic models.

  • Replicate models used to represent immiscible phase compositions (e.g., "AbPl" and "AnPl" to represent albite-rich and albite-poor compositions of the plagioclase model "Pl(h)") have been eliminated in favor of general models. Such replicates are unnecessary in Perple_X 07 because the program identifies immiscible phases automatically. Replicate models are sometimes easier to interpret and eliminate some complications associated with using a single model for immiscible phases, thus for some purposes users may wish to reintroduce replicate models.

  • Some solution models, notably Cr-, Ti-, and Mn-bearing solutions, in solution_model.dat have non-linear subdivision schemes.

• The graphical user interface PerpLith cannot be used with Perple_X 07.

• Phase equilibria can be computed explicitly as a function of the activity or fugacity of a chemical species in addition to chemical potentials as possible formerly. As a consequence the program MU_2_F used formerly to convert chemical potential diagrams to diagrams as a function of fugacity or activity is obsolete.

• The polygonalized phase diagram section strategy (Connolly and Petrini, 2002) has been eliminated. As a consequence the program POLYGON and the pseudosection tutorial are obsolete. Refer to the seismic velocity tutorial for an up-to-date illustration of how to calculate and manipulate phase diagram sections.