A Custom Geophysical Cloud Service
Access a customized 3D inversion environment to model Dias data sets
The DIAS CARAVEL cloud service provides a customized solution for geoscientists who want to maximize the value of their big data. Storage, signal processing and 3D inversion of big data each have unique hardware requirements that are not well-served by public cloud solutions.
In use for years by Dias’ geophysicists, the DIAS CARAVEL system is a proven environment for the processing and interpretation of DIAS32 and QMAGT data sets.
DIAS CARAVEL is a software application that allows users to run the full suite of SimPEG inversion codes on the purpose-built CARAVEL cloud. Utilizing modern parallelization techniques, CARAVEL is able to utilize many simultaneous compute nodes to provide results on the largest datasets in a short period of time. Inversion is often an iterative process and the ability to invert project data multiple times with varying parameters can significantly improve the final model. Available 3D codes include:
- DC resistivity & induced polarization – tensor mesh
- DC resistivity & induced polarization – octree mesh
- Full-tensor magnetic gradiometry
Our 3D world demands 3D imaging. There is a rapid industry shift to geophysical data sets with many receiver-transmitter pairs. 3D inversion of these large datasets is a challenge even with modern mesh strategies. The inter-process communication required by 3D inversion doesn’t scale well when communications have latency. For efficiency, all processes should be run on a single system rather than on multiple systems. Recent advancements provide for 128 cores on a single system. Add 2 Tb of RAM, and this becomes a powerful tool for inverting large 3D data sets.
“Data is the new science. Big Data holds the answers.”
Angela Ahrendts, Apple
At 100 times slower than ethernet, internet speed can be a pinch point for big data. By locating storage near our computer systems, we can load and save data on demand at ethernet speeds, so processing and inversions can run at full throttle.
Full-time-series capture and advanced processing algorithms demand serious processing power. Machine learning for quality control is computationally intense, and the management, stacking and formatting of big data is also challenging. These processes involve both parallelizable and single-threaded problems. A cloud that can dynamically shift loads to the most efficient architecture is needed to ensure timely results. Run your own inversion of Dias data sets without the need to purchase expensive hardware. Interact with our expert geophysical team for effective inversion strategies.