You are Home   »   News   »   View Article

Shearwater reveals 10x speed-ups with NVIDIA GPU-acceleration

Wednesday, August 30, 2023

Shearwater announces 10x speed-ups in Shearwater reveal with NVIDIA GPU-acceleration

Houston, 30th May 2023

Shearwater GeoServices, announced the achievement of a key performance milestone to its most complex seismic processing algorithms, part of the Reveal seismic processing software and based on NVIDIA GPUs to achieve 10x faster performance and improved energy efficiency This follows the announcement in August 2022 of a technology collaboration between Shearwater and NVIDIA. Seismic processing is a key workflow for energy companies to optimize oil and gas exploration and production (E&P)-providing a reliable supply of fuels to meet surging demand worldwide.

The significant step forward in these efforts is achieved through a 10x speed-up of Reverse Time Migration (RTM) and Kirchhoff algorithms, powered by NVIDIA GPUs, to lower total power consumption for compute-intensive workloads, improve energy efficiency, and reduce operating costs.

Shearwater and NVIDIA continue to collaborate to further increase and optimize the available portfolio of GPU-accelerated tools in Reveal to ensure Reveal's customers can run optimally scaling seismic workflows on their GPU-accelerated HPC infrastructure, in areas such as surface related multiple removal and FWI imaging. Enabling customers to run optimally scaled seismic workflows on their GPU-accelerated HPC infrastructure will reduce processing time and lower carbon emissions from oil and gas operations.

'This is an exciting and important collaboration and a fantastic example of the commitment by Shearwater to develop and deploy state-of-the-art technology and software for the benefit of our clients,' said Simon Telfer, SVP Software, Processing and Imaging at Shearwater. Reducing cycle time and lowering emissions are key objectives for Shearwater and, producing a 10x speedup of algorithms using power efficient GPU hardware in this partnership with NVIDIA is a significant step forwards in this work.

'Shearwater has made great progress in accelerating Reveal on NVIDIA platforms. Their close collaborations demonstrate their commitment to providing customers with the best performance at the lowest cost and carbon footprint,' said Marc Spieler, Sr. Director of Energy at NVIDIA.

Shearwater Reveal improved performance also leverages NVIDIA Bitcomp, part of the nvCOMP library used in fast lossy data compression and decompression, for reverse time migration (RTM) and was rolled out in production and used regularly across Reveal. This library allows for accelerated compression of seismic snapshots and is an efficient way to compress wave-fields. Bitcomp reduces the memory transfer load between CPU and GPU and is becoming industry standard. This results in a more efficient RTM workflow with significant speed-ups over workflows not leveraging the compression library, with one oil and gas customer achieving 8x speed-ups in application performance. Additionally, the solution allows users to run larger simulations given the available memory on the GPU.

Leading energy companies use Reveal to enhance their seismic processing workflows NVIDIA Tensor Core GPUs have been used to accelerate the performance of high-end seismic data processing and imaging algorithms in Reveal, enabling faster and more accurate estimation of subsurface properties in complex geological settings. This ultimately translates to faster time-to-oil, to improved subsurface imaging in a fixed time frame, and revenue opportunities for energy providers. In addition to E&P, Reveal is used to process seismic data for shallow hazard mapping, windfarms planning, geothermal projects, and carbon capture, utilization, and storage (CCUS).

Associated Companies
» Shearwater GeoServices
comments powered by Disqus


To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.


Latest Edition Jan-Feb 2024
Jan 2024

Download latest and back issues


Learn more about supporting Digital Energy Journal