Custom Query (454 matches)

Currently our shell script writers require a single amount of memory, but we often run jobs on mixed computing environments. For example, the SCEC queue at USC HPCC has machines with 12 and 42 GB of memory. Some calculations (ETAS) can make use of additional memory to run faster, but will run fine on lower memory systems. Our launch scripts should detect system memory reliably and adjust the JVM maximum memory statement when required (this will be optional and not the default behavior).

JFreeChart uses 0 as the basis for all ticks. This can look weird for histograms where the first bin is at, for example, 1. There should be an option to center tick labels at histogram bin centers.

Digitized notes from Morgan (because I lose pieces of paper):

doInversion() calculates rupRateSolution. Use this, along with rupList (defines which sections ruptures in which ruptures)

rupMeanMag gives magnitude for each rupture.

To make participation map: User picks Mmin and Mmax (magnitude range to plot)

For each fault section, find event rate for ruptures that include that section and are within mag range:

Participation Rate ([Mmin Mmax], i) = Sum(rupRateSolution)j

i is fault section index j is rupture index. only sum over rups when i is in (rupList)j, AND (rupMeanMag)j is within Mmin, Mmax.

Then, plot event rates for all sections (a different color for each fault section?)

Click on fault section, bring up graph of magnitude participation distribution (for entire mag range)

y axis: rate/yr x axis: mag

This is the same as plotting the par. dist. (just as calculated in the first plot) for each mag bin:

[6, 6.1], [6.1, 6.2] ... etc.