Staff Priorities - Kevin: Difference between revisions

From SCECpedia
Jump to navigationJump to search
Kmilner (talk | contribs)
No edit summary
Kmilner (talk | contribs)
No edit summary
Line 1: Line 1:
9/28/15
9/29/15


== Yesterday: ==
== Yesterday: ==
* finished MCEr code refactoring
* implemented MCEr calculations with UCERF3 based GMPE calcs
* implemented CB's proposed CyberShake/GMPE weighting algorithm:
** used Peter's deterministic UCERF3 ERF
** average UCERF3 based GMPE calculations with UCERF2 based CyberShake calcs
** issue: slow, even with slightly downsampled ERF
** 1s: full GMPE weight, 10s: full CyberShake weight. Will linearly interpolate between for now but make it easy to swap in other weighting functions
** need to precompute hazard curves on cluster, started work on that
** Apply weighted averaging directly to hazard curves (before RTGM calc) for probabilistic calcs, apply to spectrum for deterministic calculations.


== Today: ==
== Today: ==
* implement MCEr calculations with UCERF3 based GMPE calcs
* finish hazard curve precomputing (requires some modifications to cache structure)
** use Peter's deterministic UCERF3 ERF
* submit UCERF3 hazard calculations for MCEr on stampede
* if time, continue mapping of historical earthquakes to UCERF3 ruptures in order to run ETAS simulations of aftershock sequences
* investigate alternatives to Java Web Start:
** JavaFX?
** Getdown?
* figure out NaN hazard curves issue with ASK 2014


== Blocked: ==
== Blocked: ==

Revision as of 16:36, 29 September 2015

9/29/15

Yesterday:

  • implemented MCEr calculations with UCERF3 based GMPE calcs
    • used Peter's deterministic UCERF3 ERF
    • issue: slow, even with slightly downsampled ERF
    • need to precompute hazard curves on cluster, started work on that

Today:

  • finish hazard curve precomputing (requires some modifications to cache structure)
  • submit UCERF3 hazard calculations for MCEr on stampede
  • investigate alternatives to Java Web Start:
    • JavaFX?
    • Getdown?
  • figure out NaN hazard curves issue with ASK 2014

Blocked:

no

Followups:

no


Lower priority or longer term tasks

  • Fill in missing site/rupture distances in CyberShake DB for ERF 36 (200m)
  • finish mapping of historical earthquakes to UCERF3 ruptures in order to run ETAS simulations of aftershock sequences