Overview


Duration: 13 May - 24 May (10 days) 

completedin progresson holdreviewto do
393428

(vs record 61 completed sprint 19)



Epic
Story
Owner
Deliverables
Link
Validation Help Sarah/Robin as needed
Added multiple distributions to srf genValidation source perturbation (Source uncertainties)
Cybershake

Non Uniform Grid (Regional + vs30generation)

Waiting on:
Confirmed GMSimVersion

VM 2.0.2


Subduction

  • Source (SRF + Stoch)
  • Vizualisation

Jason

Started

70000 / 110,000 LF progress













Cybershake v19p5
Slurm Workflow

Task Management Changes:

  • Give the autosubmit a list of tasks that you would like to compute - defaults to IM_Calc, clean-up
  • Automatically determine if dependencies have been met and if not automatically fills them in
  • Option to call autosubmit to run just a specific single realisation
  • Autosubmit wrapper to manage "one realisation per fault" tasks e.g. VM Gen / Empirical / Rrup - with configuration file
  • On error create a new task if the number retry count of the task is > RETRY_COUNT_MAX

  • Write 2 scripts that convert the file format of LF/HF to BB. (no extra features)
  • Have alternative dependencies on IMCalc for these tasks. So IM-Calc doesn't automatically add normal BB to task list

Added new E2E test setup for 18.5.3.4 and seed sanity checking

James, Jason

1, 2) Auto_submit.py refactored to allow the user to specify the tasks to be run, and dependencies are automatically added.

3)Auto_submit.py refactored to allow the user to specify the realisations to be run, including the use of wildcards to allow multiple realisations.

6) HF and LF are able to be converted to BB without the other.

7) If HF2BB or LF2BB and IM_calc is also to be run, then they are accepted as the dependency and BB is not added.

8) Done

Task Management Refactor Plan

1, 2, 3, 6, 7) Step 6: Submit simulation jobs to HPC

2) Workflow dependencies

SeisFinder

Deploy Dev / EAP

New Empirical format

GM Selection

My Project

Folder restructure + package creation

Refactor Deagg/Hazard + Unit tests


Refactor imdb create

Viktor / Claudio / Daniel

Conceptual design of overall architecture is in progress

AWS Lambda experimented



Done

Done, currently changing again though, and might change again once interfaces are defined

Needs testing and create PR

SeisFinder Architecture

AWS services : EC2, S3 and Lambda


New empirical data format : < 10% of the original size (~200Gb)

TestEmpirical tests


Ground failure

Generate lots of data

James/Jason



IM_Calc
  • Vahid's integration for Advanced IMs

Refactor of source to site distance calculations

Jason

Claudio



Create PR, otherwise done


Bug fixes



MiscHelp Karim generate hazard as required


 



  • No labels