2.1.1 Distributed Computing & Parallelization

To speed up processing time OpendTect supports batch processing of attribute volumes and other computer intensive processes. The user has the option to run batch processes on all computers (Linux, Mac or Windows operated workstations and clusters) in the network. Distributed computing is highly recommended in operational settings. Corporate users are advised to contact their system administrator to enable multi-machine processing. Please see the System Administrator's Documentation for details.

In addition to distributed computing OpendTect also supports multi-threading (parallel processing) of many algorithms. A thread of execution is the smallest unit of processing that can be scheduled by an operating system. It generally results from a fork of a computer program into two or more concurrently running tasks. The implementation of threads and processes differs from one operating system to another, but in most cases, a thread is contained inside a process.

Multi-threading allows multiple threads to exist within the context of a single or multiple process. These threads share the process' resources (such as memory) but are able to execute independently.

OpendTect uses the power of multi-threading to run different applications in data loading, visualization, processing etc...

Processes that use Multi-threading

Dip-Steering Algorithms

Compute a Steering Cube with BG Steering

Completely re-designed in Nov. 2015

Compute a Steering Cube with FFT

Completely re-designed in Nov. 2015

Apply Full Steering dip-steering in attributes and filters

 

Apply Central Steering in attributes and filters

 

Attributes

Convolve (all except Wavelet option)

 

Curvature

 

Dip Angle

 

Velocity Fan Filter (=DipFilter)

 

Energy (all except Gradient option)

 

Event

 

Frequency Filter

multi-threading implemented in March 2014

Frequency

multi-threading implemented in March 2014

HorizonCube Data

 

HorizonCube Density

 

HorizonCube Layer

 

Hilbert

 

Instantaneous

multi-threading improved in March 2014

Local Fluid Contact Finder

 

Maths (except when expression is recursive)

 

Polar Dip

 

Position

 

PreStack

 

Reference

 

Scaling (all except scaling type AGC and stats type = detrend)

 

Semblance

 

Similarity

 

Spectral Decomposition

multi-threading implemented in March 2014

Texture

 

Tutorial

 

Volume Statistics

 

Processes that do not (yet) use multi-threading

Attributes

Constant Steering

Convolve (Wavelet option)

DeltaResample

Energy (Gradient option)

Event Steering

FaultDip

FingerPrint

GapDecon

HorizonCube Curvature

HorizonCube Dip

HorizonCube Spacing

Horizon

Match Delta

Maths (recursive expression)

Perpendicular Dip Extractor

SampleValue

Scaling (scaling type AGC and stats type = detrend)

Shift

*multi-threading implemented in March 2014
**multi-threading improved in March 2014