The code computes a sum of damped, decaying, or periodic oscillators, convolved
with (unnormalized) Gaussians, on a grid. The analysis computes the sum of all
voxel-wise autocorrelations for all window sizes up to the specified input (via
-w
option). It also finds the voxels with the k
strongest autocorrelations
(for every shift).
The oscillators' locations and parameters are specified in an input file. See
inputs folder for examples. The generate_input.py
script can generate a set
of randomly initialized oscillators.
The simulation code is in oscillator.cpp
.
The analysis interface is specified in analysis.h
.
The actual analysis code is in analysis.cpp
.
To run:
mpirun -n ... ./oscillator sample.osc
Usage (ENABLE_SENSEI=ON):
./bin/oscillator [OPTIONS] OSCILLATORS.txt
Options:
-b, --blocks INT number of blocks to use [default: 1]
-s, --shape POINT domain shape [default: 64 64 64]
-t, --dt FLOAT time step [default: 0.01]
-f, --config STRING SENSEI analysis configuration xml (required)
--t-end FLOAT end time [default: 10]
--sync synchronize after each time step
-h, --help show help
Usage (ENABLE_SENSEI=OFF):
./bin/oscillator [OPTIONS] OSCILLATORS.txt
Options:
-b, --blocks INT number of blocks to use [default: 1]
-s, --shape POINT domain shape [default: 64 64 64]
-t, --dt FLOAT time step [default: 0.01]
-w, --window UNSIGNED LONG analysis window [default: 10]
-k, --k-max UNSIGNED LONG number of strongest autocorrelations to report [default: 3]
--t-end FLOAT end time [default: 10]
--sync synchronize after each time step
-h, --help show help