+ All Categories
Home > Documents > Improving Earthquake Forecasts using USC HPCC

Improving Earthquake Forecasts using USC HPCC

Date post: 24-Feb-2016
Category:
Upload: lixue
View: 36 times
Download: 0 times
Share this document with a friend
Description:
Improving Earthquake Forecasts using USC HPCC. Scott Callaghan Southern California Earthquake Center University of Southern California SC12. What are earthquake forecasts?. Want to describe possible earthquakes in a region Seismic hazard maps Insurance rates Building codes - PowerPoint PPT Presentation
19
Improving Earthquake Forecasts using USC HPCC Scott Callaghan Southern California Earthquake Center University of Southern California SC12
Transcript
Page 1: Improving Earthquake Forecasts using USC HPCC

Improving Earthquake Forecasts using USC HPCC

Scott CallaghanSouthern California Earthquake CenterUniversity of Southern CaliforniaSC12

Page 2: Improving Earthquake Forecasts using USC HPCC

What are earthquake forecasts?

• Want to describe possible earthquakes in a region– Seismic hazard maps– Insurance rates– Building codes

• Determine faults,magnitudes,earthquake rates

• Forecast producedevery 5 yearsfor California

Page 3: Improving Earthquake Forecasts using USC HPCC

Components of earthquake forecasts

• Integrate data from many sources– Magnitude-frequency distribution– Paleoseismicity– Slip rates– Sanity checks

• Try to satisfyconstraintsas closelyas possible

Page 4: Improving Earthquake Forecasts using USC HPCC

The Grand Inversion

• Divide faults into small segments• Consider 1 or more segments together• Solve for rates, given constraints

– 234k rates– 30k constraints

• Minimize error• Need to run

thousands of times– Underdetermined– Multiple branches

Page 5: Improving Earthquake Forecasts using USC HPCC

Simulated Annealing (SA)

• Iterative approach for solving optimization problems• Works by reducing ‘energy’ heuristic

– Calculate energy of current state– Calculate energy of neighboring state– Probability of moving to a neighbor state is proportional to

the temperature and inversely proportional to the energy difference• If energy is less, always move• If energy is greater, occasionally move

– Over time, reduce temperature to converge on a minimum energy (not necessarily global minimum)

• Serial version too slow for Grand Inversion

Page 6: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated Annealing

• Needed to parallelize algorithm• Have each core perform serial SA for some number

of iterations (nSubIterations)• Share best answer with all cores• Continue until stopping criteria is met• Able to cover more search space quickly

Page 7: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated Annealing

Page 8: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated AnnealingnNodes = 5

Page 9: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated AnnealingnSubIterations = 200

Page 10: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated Annealing

Page 11: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated Annealing

Page 12: Improving Earthquake Forecasts using USC HPCC

Parallel Simulated Annealing

Page 13: Improving Earthquake Forecasts using USC HPCC

Implementation

• Seismology code– OpenSHA - http://www.opensha.org, open source– Java-based, object-oriented

• Both Java MPI and threaded versions• Why Java?

– Codebase in Java, avoid cost of porting– Scientists already comfortable with Java– Avoid maintaining separate serial and parallel codebases

• OpenSHA is used in many other applications

Page 14: Improving Earthquake Forecasts using USC HPCC

Performance

LegendSingle Node (thin lines)• 1 thread• 2 threads• 4 threads• 8 threads

Multiple Nodes (4 threads ea.)• 2 nodes (8 threads)• 5 nodes (20

threads)• 10 nodes (40

threads)• 20 nodes (80

threads)• 50 nodes (200

threads)• 100 nodes (400

threads)

Page 15: Improving Earthquake Forecasts using USC HPCC

Optimal

Actual

Sqrt(threads)

Page 16: Improving Earthquake Forecasts using USC HPCC
Page 17: Improving Earthquake Forecasts using USC HPCC

Optimization

• Improve performance of serial section• Energy calculation:

– [1x234000] x [234000x30000] = [1x30000]– Calculate misfits

• Switched to Parallel Colt• Stopped performing entire matrix multiplication each

iteration– Only compute differences, not whole matrix– 100x speedup– Additional 10x with caching

Page 18: Improving Earthquake Forecasts using USC HPCC

HPCC runs

• 7682 inversions• Best use of SUs is 1 node per inversion

– 5 hours on 8 cores/node per inversion• Continuing to run inversions with new data and

models• Quick way to take new inputs and determine impact

on rates

Page 19: Improving Earthquake Forecasts using USC HPCC

Questions?


Recommended