C405 Bad
C408 Did I send in or forget?
C394/C395 Didn't Tony Fill in
C425 Resubmit -- Haupt?

Low 12.11
Medium 13.6
High 15.14 plus insurance

GEM Meeting ***********************************
Mora: Building an earthquake PSE -- C++ Linux

Japan: Also a PSE Rupture -- Crustal Activity(Data) -- Core-Mantle
       40 Teraflops
       
China: LNM Institute (Lab of Nonlinear Mechanics) collaboration involves BUAA
       SMDM Statistical Mesoscopic Damage Mechanics
       LURR Loading Unloading Y isgood precursor
       Hope to predict 65% accuray for months-- year from now if past 2-5 years

Gell-Mann **********
Complexity:
Computational Complexity
Algorithm Information Complexity
2 parts -- regular plus incidental/statistical

Effectivity Complexity is regular part
Frozen Accidents of branching determine effective complexity

Plectics: Simplicity plus Complexity
Simple simulations: What do they mean?

the Quark and the Jaguar Book

Crypticity Charlie Bennett -- how hard is it find complexity

Note Pseudo Random Numbers
Turbulence

Turcotte *************
Chaos
Lorenz

Fragmentation
Los Alamos failed to calculate
Large computations not succesful

MGM: Use computations as experimental data NOT as a crutch

"Power Laws" are not an acceptable "statistical law"!

Maximize Ignorance
Fix first and second moments gives Exponential and Gaussian
Sorin Solomon paper "Power law is just a log of a exponential"

************************************************************
Minster/Jordan ******
Hazard Mitigation

What can you do ******************************
Event Specific Prediction
Probabilistic forecasting of large erathquakes
Ground Motion from "typical expected earthquakes"

Issues *********************
Difficult Physics
Difficult Observations
Lack of well funded effort
Lack of qualified people

Issues in Earthquake Simulation
CPU Cycles
Simulate Systems (like Sand Dunes) not amenable to PDE
New Algorithms
Mining databases
Visualization
Collaboration and Information Sharing

EarthScope
Multi Sensor Multi Agency 

SCEC
  - Multidisciplinary Success
ACES
  - Working Group
GEM
  - Interaction with ITR Community

Tullis ********************************
Roughly 10 meter resolution simulation
Full 3D Forward Problems
Jackson: Detailed Deterministic Model

Online Data ***************************
Legal Issues seem a problem
copyright databases (already passed in Europe)

MHPCC
Classified Computing 
New System in 2000 getting total of 0.75 Teraflop
6 Terabyte Disk
Must get a NAC to access 

Issues
What does one need?
Ideas 

Computers -- What size when
 - Parallel ?
 - Shared or Distributed (larger)

Application Area
 - 3D Simulations
 - "Intelligent" Datamining (data from nature or model)
 - Crisis Management/Training
 
Application Software 
 - Individual or Group (as in Japan)
 - Qualitative Structure "Irregular FEM" "Fast Multipole"
Systems and Tools Software
 - MPI
 - Grid Generation
 - Multipole
 - Visualization
Data Storage and Access
Education and Training
 - Parallel Computing
 - Training -- Teach MPI
 - Education
Systems Integration
 - PSE/Toolkit (web-based portal)
 - (XML) Standards for data / program interfaces

Can one integrate US and Japanses/Australian "programs"
Is it Innovative
Is there anything special needed to be done or will 
   MHPCC/Ames/PACI suffice

Data Explorer Free
Computational Steering 
Immersive 
PDE Solvers

Group Approach
1) Rupture Dynamics -- link waves in bulk; complex boundaries
(lot of Jazpan/Australia Interest)
2) NASA Data Support
3) Real Time Issues

Mora wants "plug and play" bulk/surface codes
Lessons from other fields? (Minster)

Australia/Japan collaboration including code linkage
at boundaries

Unclear how to set up "Steve Day" problem -- is it 
ready to go yet. Doubt it

Bosl: Component Technology 

Can we define interface between volume and boundary
simulations (boundary = fault)

SCEC ************ 3 year proposal **********************
1) Micro Scale
   Calculate Friction
2) Meso Scale
   Rupture Dynamics and link to Earthquake Dynamics
   Scenarios
   What about CMU Group?
3) Macro Scale
   Large Systems

What about IT? Education (Jordan likes but
others not much interest)

Mora: doesn't agree - believes in more ambitious program

Japan: $600M Computer over 5 years
        $30M Software over 5 years

China $100 million yuan over ? years

SCEC $5.5M year 
Deliberatley disorganized approach to Science