On Sun, Feb 10, 2013 at 1:05 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:
> > On Sat, Feb 9, 2013 at 8:04 PM, Matthew Knepley <knepley at gmail.com>wrote: >> >> What's your plan for implementing this? >>> >> >> My plan was to try the easiest thing I could think of first: >> >> a) Make a replacement monitor for SNES and KSP >> >> It would have a big honking sentinel before each solve and >> iterations in Python array format >> >> b) Run a braindead Python parser over this which snips out the monitors >> >> c) Diff the monitor sections with a numerical tolerance >> >> d) Diff the whole thing, and throw out differences in the monitor >> sections, since we know the line numbers. >> Maybe it would be easier to just remove those lines, but we need to >> experiment a little. >> > > Okay, so this would all be implemented in C? > Does the Bear(y) shit in the woods? > If you want Python array format, you may as well use JSON since it's more > of an exchange standard and it's the most convenient format for monitoring > in a web-based GUI. There are a bunch of good JSON libraries, ranging from > the ultra-compact http://zserge.bitbucket.org/jsmn.html and > http://ccodearchive.net/info/json.html, to the object-oriented > http://www.digip.org/jansson/doc/dev/tutorial.html. > The phrase "may as well" is very slippery. Python arrays are just [x, y, z], whereas JSON needs a library. However, the format is is now way crucial to the scheme. I want to get the workflow up an running with the simplest format. We will inevitably tweak it was we add tons of crap on top if it works. Matt -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130210/092a0fef/attachment.html>
