With the exponential growth of petroleum exploration data,the data computation and disk storage of various application software systems have also increased,posing great challenges to traditional parallel processing architectures.This paper proposes the design and optimization scheme of a parallel processing system architecture for massive seismic data.By analyzing the sensitivity of each software module to hardware parameters such as CPU,memory,local disk and network transmission during seismic data processing,as well as optimizing the configuration of NFS,RPC and Buffer,the efficient operation of the cluster has been achieved.Actual production applications have shown that the system has unique advantages in supporting parallel processing of massive data at the level of 100 TB,as well as compatibility issues with multiple operating system software,which strongly supports the development of seismic data processing business.
维普
万方数据